Hide ToC

Using Strimzi (0.24.0)

Table of Contents

1. Overview of Strimzi

Strimzi simplifies the process of running Apache Kafka in a Kubernetes cluster.

This guide provides instructions for configuring Kafka components and using Strimzi Operators. Procedures relate to how you might want to modify your deployment and introduce additional features, such as Cruise Control or distributed tracing.

You can configure your deployment using Strimzi custom resources. The Custom resource API reference describes the properties you can use in your configuration.

Note
Looking to get started with Strimzi? For step-by-step deployment instructions, see the Deploying and Upgrading Strimzi guide.

1.1. Kafka capabilities

The underlying data stream-processing capabilities and component architecture of Kafka can deliver:

  • Microservices and other applications to share data with extremely high throughput and low latency

  • Message ordering guarantees

  • Message rewind/replay from data storage to reconstruct an application state

  • Message compaction to remove old records when using a key-value log

  • Horizontal scalability in a cluster configuration

  • Replication of data to control fault tolerance

  • Retention of high volumes of data for immediate access

1.2. Kafka use cases

Kafka’s capabilities make it suitable for:

  • Event-driven architectures

  • Event sourcing to capture changes to the state of an application as a log of events

  • Message brokering

  • Website activity tracking

  • Operational monitoring through metrics

  • Log collection and aggregation

  • Commit logs for distributed systems

  • Stream processing so that applications can respond to data in real time

1.3. How Strimzi supports Kafka

Strimzi provides container images and Operators for running Kafka on Kubernetes. Strimzi Operators are fundamental to the running of Strimzi. The Operators provided with Strimzi are purpose-built with specialist operational knowledge to effectively manage Kafka.

Operators simplify the process of:

  • Deploying and running Kafka clusters

  • Deploying and running Kafka components

  • Configuring access to Kafka

  • Securing access to Kafka

  • Upgrading Kafka

  • Managing brokers

  • Creating and managing topics

  • Creating and managing users

1.4. Strimzi Operators

Strimzi supports Kafka using Operators to deploy and manage the components and dependencies of Kafka to Kubernetes.

Operators are a method of packaging, deploying, and managing a Kubernetes application. Strimzi Operators extend Kubernetes functionality, automating common and complex tasks related to a Kafka deployment. By implementing knowledge of Kafka operations in code, Kafka administration tasks are simplified and require less manual intervention.

Operators

Strimzi provides Operators for managing a Kafka cluster running within a Kubernetes cluster.

Cluster Operator

Deploys and manages Apache Kafka clusters, Kafka Connect, Kafka MirrorMaker, Kafka Bridge, Kafka Exporter, and the Entity Operator

Entity Operator

Comprises the Topic Operator and User Operator

Topic Operator

Manages Kafka topics

User Operator

Manages Kafka users

The Cluster Operator can deploy the Topic Operator and User Operator as part of an Entity Operator configuration at the same time as a Kafka cluster.

Operators within the Strimzi architecture

Operators within the Strimzi architecture

1.4.1. Cluster Operator

Strimzi uses the Cluster Operator to deploy and manage clusters for:

  • Kafka (including ZooKeeper, Entity Operator, Kafka Exporter, and Cruise Control)

  • Kafka Connect

  • Kafka MirrorMaker

  • Kafka Bridge

Custom resources are used to deploy the clusters.

For example, to deploy a Kafka cluster:

  • A Kafka resource with the cluster configuration is created within the Kubernetes cluster.

  • The Cluster Operator deploys a corresponding Kafka cluster, based on what is declared in the Kafka resource.

The Cluster Operator can also deploy (through configuration of the Kafka resource):

  • A Topic Operator to provide operator-style topic management through KafkaTopic custom resources

  • A User Operator to provide operator-style user management through KafkaUser custom resources

The Topic Operator and User Operator function within the Entity Operator on deployment.

Example architecture for the Cluster Operator

The Cluster Operator creates and deploys Kafka and ZooKeeper clusters

1.4.2. Topic Operator

The Topic Operator provides a way of managing topics in a Kafka cluster through Kubernetes resources.

Example architecture for the Topic Operator

The Topic Operator manages topics for a Kafka cluster via KafkaTopic resources

The role of the Topic Operator is to keep a set of KafkaTopic Kubernetes resources describing Kafka topics in-sync with corresponding Kafka topics.

Specifically, if a KafkaTopic is:

  • Created, the Topic Operator creates the topic

  • Deleted, the Topic Operator deletes the topic

  • Changed, the Topic Operator updates the topic

Working in the other direction, if a topic is:

  • Created within the Kafka cluster, the Operator creates a KafkaTopic

  • Deleted from the Kafka cluster, the Operator deletes the KafkaTopic

  • Changed in the Kafka cluster, the Operator updates the KafkaTopic

This allows you to declare a KafkaTopic as part of your application’s deployment and the Topic Operator will take care of creating the topic for you. Your application just needs to deal with producing or consuming from the necessary topics.

The Topic Operator maintains information about each topic in a topic store, which is continually synchronized with updates from Kafka topics or Kubernetes KafkaTopic custom resources. Updates from operations applied to a local in-memory topic store are persisted to a backup topic store on disk. If a topic is reconfigured or reassigned to other brokers, the KafkaTopic will always be up to date.

1.4.3. User Operator

The User Operator manages Kafka users for a Kafka cluster by watching for KafkaUser resources that describe Kafka users, and ensuring that they are configured properly in the Kafka cluster.

For example, if a KafkaUser is:

  • Created, the User Operator creates the user it describes

  • Deleted, the User Operator deletes the user it describes

  • Changed, the User Operator updates the user it describes

Unlike the Topic Operator, the User Operator does not sync any changes from the Kafka cluster with the Kubernetes resources. Kafka topics can be created by applications directly in Kafka, but it is not expected that the users will be managed directly in the Kafka cluster in parallel with the User Operator.

The User Operator allows you to declare a KafkaUser resource as part of your application’s deployment. You can specify the authentication and authorization mechanism for the user. You can also configure user quotas that control usage of Kafka resources to ensure, for example, that a user does not monopolize access to a broker.

When the user is created, the user credentials are created in a Secret. Your application needs to use the user and its credentials for authentication and to produce or consume messages.

In addition to managing credentials for authentication, the User Operator also manages authorization rules by including a description of the user’s access rights in the KafkaUser declaration.

1.4.4. Feature gates in Strimzi Operators

You can enable and disable some features of operators using feature gates.

Feature gates are set in the operator configuration and have three stages of maturity: alpha, beta, or General Availability (GA).

For more information, see Feature gates.

1.5. Strimzi custom resources

A deployment of Kafka components to a Kubernetes cluster using Strimzi is highly configurable through the application of custom resources. Custom resources are created as instances of APIs added by Custom resource definitions (CRDs) to extend Kubernetes resources.

CRDs act as configuration instructions to describe the custom resources in a Kubernetes cluster, and are provided with Strimzi for each Kafka component used in a deployment, as well as users and topics. CRDs and custom resources are defined as YAML files. Example YAML files are provided with the Strimzi distribution.

CRDs also allow Strimzi resources to benefit from native Kubernetes features like CLI accessibility and configuration validation.

1.5.1. Strimzi custom resource example

CRDs require a one-time installation in a cluster to define the schemas used to instantiate and manage Strimzi-specific resources.

After a new custom resource type is added to your cluster by installing a CRD, you can create instances of the resource based on its specification.

Depending on the cluster setup, installation typically requires cluster admin privileges.

Note
Access to manage custom resources is limited to Strimzi administrators. For more information, see Designating Strimzi administrators in the Deploying and Upgrading Strimzi guide.

A CRD defines a new kind of resource, such as kind:Kafka, within a Kubernetes cluster.

The Kubernetes API server allows custom resources to be created based on the kind and understands from the CRD how to validate and store the custom resource when it is added to the Kubernetes cluster.

Warning
When CRDs are deleted, custom resources of that type are also deleted. Additionally, the resources created by the custom resource, such as pods and statefulsets are also deleted.

Each Strimzi-specific custom resource conforms to the schema defined by the CRD for the resource’s kind. The custom resources for Strimzi components have common configuration properties, which are defined under spec.

To understand the relationship between a CRD and a custom resource, let’s look at a sample of the CRD for a Kafka topic.

Kafka topic CRD
apiVersion: kafka.strimzi.io/v1beta2
kind: CustomResourceDefinition
metadata: (1)
  name: kafkatopics.kafka.strimzi.io
  labels:
    app: strimzi
spec: (2)
  group: kafka.strimzi.io
  versions:
    v1beta2
  scope: Namespaced
  names:
    # ...
    singular: kafkatopic
    plural: kafkatopics
    shortNames:
    - kt (3)
  additionalPrinterColumns: (4)
      # ...
  subresources:
    status: {} (5)
  validation: (6)
    openAPIV3Schema:
      properties:
        spec:
          type: object
          properties:
            partitions:
              type: integer
              minimum: 1
            replicas:
              type: integer
              minimum: 1
              maximum: 32767
      # ...
  1. The metadata for the topic CRD, its name and a label to identify the CRD.

  2. The specification for this CRD, including the group (domain) name, the plural name and the supported schema version, which are used in the URL to access the API of the topic. The other names are used to identify instance resources in the CLI. For example, kubectl get kafkatopic my-topic or kubectl get kafkatopics.

  3. The shortname can be used in CLI commands. For example, kubectl get kt can be used as an abbreviation instead of kubectl get kafkatopic.

  4. The information presented when using a get command on the custom resource.

  5. The current status of the CRD as described in the schema reference for the resource.

  6. openAPIV3Schema validation provides validation for the creation of topic custom resources. For example, a topic requires at least one partition and one replica.

Note
You can identify the CRD YAML files supplied with the Strimzi installation files, because the file names contain an index number followed by ‘Crd’.

Here is a corresponding example of a KafkaTopic custom resource.

Kafka topic custom resource
apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaTopic (1)
metadata:
  name: my-topic
  labels:
    strimzi.io/cluster: my-cluster (2)
spec: (3)
  partitions: 1
  replicas: 1
  config:
    retention.ms: 7200000
    segment.bytes: 1073741824
status:
  conditions: (4)
    lastTransitionTime: "2019-08-20T11:37:00.706Z"
    status: "True"
    type: Ready
  observedGeneration: 1
  / ...
  1. The kind and apiVersion identify the CRD of which the custom resource is an instance.

  2. A label, applicable only to KafkaTopic and KafkaUser resources, that defines the name of the Kafka cluster (which is same as the name of the Kafka resource) to which a topic or user belongs.

  3. The spec shows the number of partitions and replicas for the topic as well as the configuration parameters for the topic itself. In this example, the retention period for a message to remain in the topic and the segment file size for the log are specified.

  4. Status conditions for the KafkaTopic resource. The type condition changed to Ready at the lastTransitionTime.

Custom resources can be applied to a cluster through the platform CLI. When the custom resource is created, it uses the same validation as the built-in resources of the Kubernetes API.

After a KafkaTopic custom resource is created, the Topic Operator is notified and corresponding Kafka topics are created in Strimzi.

1.6. Listener configuration

Listeners are used to connect to Kafka brokers.

Strimzi provides a generic GenericKafkaListener schema with properties to configure listeners through the Kafka resource.

The GenericKafkaListener provides a flexible approach to listener configuration. You can specify properties to configure internal listeners for connecting within the Kubernetes cluster, or external listeners for connecting outside the Kubernetes cluster.

Each listener is defined as an array in the Kafka resource. You can configure as many listeners as required, as long as their names and ports are unique.

You might want to configure multiple external listeners, for example, to handle access from networks that require different authentication mechanisms. Or you might need to join your Kubernetes network to an outside network. In which case, you can configure internal listeners (using the useServiceDnsDomain property) so that the Kubernetes service DNS domain (typically .cluster.local) is not used.

For more information on the configuration options available for listeners, see the GenericKafkaListener schema reference.

Configuring listeners to secure access to Kafka brokers

You can configure listeners for secure connection using authentication. For more information on securing access to Kafka brokers, see Managing access to Kafka.

Configuring external listeners for client access outside Kubernetes

You can configure external listeners for client access outside a Kubernetes environment using a specified connection mechanism, such as a loadbalancer. For more information on the configuration options for connecting an external client, see Configuring external listeners.

Listener certificates

You can provide your own server certificates, called Kafka listener certificates, for TLS listeners or external listeners which have TLS encryption enabled. For more information, see Kafka listener certificates.

1.7. Document Conventions

Replaceables

In this document, replaceable text is styled in monospace, with italics, uppercase, and hyphens.

For example, in the following code, you will want to replace MY-NAMESPACE with the name of your namespace:

sed -i 's/namespace: .*/namespace: MY-NAMESPACE/' install/cluster-operator/*RoleBinding*.yaml

2. Configuring a Strimzi deployment

This chapter describes how to configure different aspects of the supported deployments using custom resources:

  • Kafka clusters

  • Kafka Connect clusters

  • Kafka Connect clusters with Source2Image support

  • Kafka MirrorMaker

  • Kafka Bridge

  • Cruise Control

Note
Labels applied to a custom resource are also applied to the Kubernetes resources comprising Kafka MirrorMaker. This provides a convenient mechanism for resources to be labeled as required.

The Deploying and Upgrading Strimzi guide describes how to monitor your Strimzi deployment.

2.1. Kafka cluster configuration

This section describes how to configure a Kafka deployment in your Strimzi cluster. A Kafka cluster is deployed with a ZooKeeper cluster. The deployment can also include the Topic Operator and User Operator, which manage Kafka topics and users.

You configure Kafka using the Kafka resource. Configuration options are also available for ZooKeeper and the Entity Operator within the Kafka resource. The Entity Operator comprises the Topic Operator and User Operator.

The full schema of the Kafka resource is described in the Kafka schema reference.

Listener configuration

You configure listeners for connecting clients to Kafka brokers. For more information on configuring listeners for connecting brokers, see Listener configuration.

Authorizing access to Kafka

You can configure your Kafka cluster to allow or decline actions executed by users. For more information on securing access to Kafka brokers, see Managing access to Kafka.

Managing TLS certificates

When deploying Kafka, the Cluster Operator automatically sets up and renews TLS certificates to enable encryption and authentication within your cluster. If required, you can manually renew the cluster and client CA certificates before their renewal period ends. You can also replace the keys used by the cluster and client CA certificates. For more information, see Renewing CA certificates manually and Replacing private keys.

Additional resources

2.1.1. Configuring Kafka

Use the properties of the Kafka resource to configure your Kafka deployment.

As well as configuring Kafka, you can add configuration for ZooKeeper and the Strimzi Operators. Common configuration properties, such as logging and healthchecks, are configured independently for each component.

This procedure shows only some of the possible configuration options, but those that are particularly important include:

  • Resource requests (CPU / Memory)

  • JVM options for maximum and minimum memory allocation

  • Listeners (and authentication of clients)

  • Authentication

  • Storage

  • Rack awareness

  • Metrics

  • Cruise Control for cluster rebalancing

Kafka versions

The log.message.format.version and inter.broker.protocol.version properties for the Kafka config must be the versions supported by the specified Kafka version (spec.kafka.version). The properties represent the log format version appended to messages and the version of Kafka protocol used in a Kafka cluster. Updates to these properties are required when upgrading your Kafka version. For more information, see Upgrading Kafka in the Deploying and Upgrading Strimzi guide.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

See the Deploying and Upgrading Strimzi guide for instructions on deploying a:

Procedure
  1. Edit the spec properties for the Kafka resource.

    The properties you can configure are shown in this example configuration:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    metadata:
      name: my-cluster
    spec:
      kafka:
        replicas: 3 (1)
        version: 2.8.0 (2)
        logging: (3)
          type: inline
          loggers:
            kafka.root.logger.level: "INFO"
        resources: (4)
          requests:
            memory: 64Gi
            cpu: "8"
          limits:
            memory: 64Gi
            cpu: "12"
        readinessProbe: (5)
          initialDelaySeconds: 15
          timeoutSeconds: 5
        livenessProbe:
          initialDelaySeconds: 15
          timeoutSeconds: 5
        jvmOptions: (6)
          -Xms: 8192m
          -Xmx: 8192m
        image: my-org/my-image:latest (7)
        listeners: (8)
          - name: plain (9)
            port: 9092 (10)
            type: internal (11)
            tls: false (12)
            configuration:
              useServiceDnsDomain: true (13)
          - name: tls
            port: 9093
            type: internal
            tls: true
            authentication: (14)
              type: tls
          - name: external (15)
            port: 9094
            type: route
            tls: true
            configuration:
              brokerCertChainAndKey: (16)
                secretName: my-secret
                certificate: my-certificate.crt
                key: my-key.key
        authorization: (17)
          type: simple
        config: (18)
          auto.create.topics.enable: "false"
          offsets.topic.replication.factor: 3
          transaction.state.log.replication.factor: 3
          transaction.state.log.min.isr: 2
          log.message.format.version: 2.8
          inter.broker.protocol.version: 2.8
          ssl.cipher.suites: "TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384" (19)
          ssl.enabled.protocols: "TLSv1.2"
          ssl.protocol: "TLSv1.2"
        storage: (20)
          type: persistent-claim (21)
          size: 10000Gi (22)
        rack: (23)
          topologyKey: topology.kubernetes.io/zone
        metricsConfig: (24)
          type: jmxPrometheusExporter
          valueFrom:
            configMapKeyRef: (25)
              name: my-config-map
              key: my-key
        # ...
      zookeeper: (26)
        replicas: 3 (27)
        logging: (28)
          type: inline
          loggers:
            zookeeper.root.logger: "INFO"
        resources:
          requests:
            memory: 8Gi
            cpu: "2"
          limits:
            memory: 8Gi
            cpu: "2"
        jvmOptions:
          -Xms: 4096m
          -Xmx: 4096m
        storage:
          type: persistent-claim
          size: 1000Gi
        metricsConfig:
          # ...
      entityOperator: (29)
        tlsSidecar: (30)
          resources:
            requests:
              cpu: 200m
              memory: 64Mi
            limits:
              cpu: 500m
              memory: 128Mi
        topicOperator:
          watchedNamespace: my-topic-namespace
          reconciliationIntervalSeconds: 60
          logging: (31)
            type: inline
            loggers:
              rootLogger.level: "INFO"
          resources:
            requests:
              memory: 512Mi
              cpu: "1"
            limits:
              memory: 512Mi
              cpu: "1"
        userOperator:
          watchedNamespace: my-topic-namespace
          reconciliationIntervalSeconds: 60
          logging: (32)
            type: inline
            loggers:
              rootLogger.level: INFO
          resources:
            requests:
              memory: 512Mi
              cpu: "1"
            limits:
              memory: 512Mi
              cpu: "1"
      kafkaExporter: (33)
        # ...
      cruiseControl: (34)
        # ...
        tlsSidecar: (35)
        # ...
    1. The number of replica nodes. If your cluster already has topics defined, you can scale clusters.

    2. Kafka version, which can be changed to a supported version by following the upgrade procedure.

    3. Specified Kafka loggers and log levels added directly (inline) or indirectly (external) through a ConfigMap. A custom ConfigMap must be placed under the log4j.properties key. For the Kafka kafka.root.logger.level logger, you can set the log level to INFO, ERROR, WARN, TRACE, DEBUG, FATAL or OFF.

    4. Requests for reservation of supported resources, currently cpu and memory, and limits to specify the maximum resources that can be consumed.

    5. Healthchecks to know when to restart a container (liveness) and when a container can accept traffic (readiness).

    6. JVM configuration options to optimize performance for the Virtual Machine (VM) running Kafka.

    7. ADVANCED OPTION: Container image configuration, which is recommended only in special situations.

    8. Listeners configure how clients connect to the Kafka cluster via bootstrap addresses. Listeners are configured as internal or external listeners for connection from inside or outside the Kubernetes cluster.

    9. Name to identify the listener. Must be unique within the Kafka cluster.

    10. Port number used by the listener inside Kafka. The port number has to be unique within a given Kafka cluster. Allowed port numbers are 9092 and higher with the exception of ports 9404 and 9999, which are already used for Prometheus and JMX. Depending on the listener type, the port number might not be the same as the port number that connects Kafka clients.

    11. Listener type specified as internal, or for external listeners, as route, loadbalancer, nodeport or ingress.

    12. Enables TLS encryption for each listener. Default is false. TLS encryption is not required for route listeners.

    13. Defines whether the fully-qualified DNS names including the cluster service suffix (usually .cluster.local) are assigned.

    14. Listener authentication mechanism specified as mutual TLS, SCRAM-SHA-512 or token-based OAuth 2.0.

    15. External listener configuration specifies how the Kafka cluster is exposed outside Kubernetes, such as through a route, loadbalancer or nodeport.

    16. Optional configuration for a Kafka listener certificate managed by an external Certificate Authority. The brokerCertChainAndKey specifies a Secret that contains a server certificate and a private key. You can configure Kafka listener certificates on any listener with enabled TLS encryption.

    17. Authorization enables simple, OAUTH 2.0, or OPA authorization on the Kafka broker. Simple authorization uses the AclAuthorizer Kafka plugin.

    18. The config specifies the broker configuration. Standard Apache Kafka configuration may be provided, restricted to those properties not managed directly by Strimzi.

    19. SSL properties for listeners with TLS encryption enabled to enable a specific cipher suite or TLS version.

    20. Storage is configured as ephemeral, persistent-claim or jbod.

    21. Storage size for persistent volumes may be increased and additional volumes may be added to JBOD storage.

    22. Persistent storage has additional configuration options, such as a storage id and class for dynamic volume provisioning.

    23. Rack awareness is configured to spread replicas across different racks. A topologykey must match the label of a cluster node.

    24. Prometheus metrics enabled. In this example, metrics are configured for the Prometheus JMX Exporter (the default metrics exporter).

    25. Prometheus rules for exporting metrics to a Grafana dashboard through the Prometheus JMX Exporter, which are enabled by referencing a ConfigMap containing configuration for the Prometheus JMX exporter. You can enable metrics without further configuration using a reference to a ConfigMap containing an empty file under metricsConfig.valueFrom.configMapKeyRef.key.

    26. ZooKeeper-specific configuration, which contains properties similar to the Kafka configuration.

    27. The number of ZooKeeper nodes. ZooKeeper clusters or ensembles usually run with an odd number of nodes, typically three, five, or seven. The majority of nodes must be available in order to maintain an effective quorum. If the ZooKeeper cluster loses its quorum, it will stop responding to clients and the Kafka brokers will stop working. Having a stable and highly available ZooKeeper cluster is crucial for Strimzi.

    28. Specified ZooKeeper loggers and log levels.

    29. Entity Operator configuration, which specifies the configuration for the Topic Operator and User Operator.

    30. Entity Operator TLS sidecar configuration. Entity Operator uses the TLS sidecar for secure communication with ZooKeeper.

    31. Specified Topic Operator loggers and log levels. This example uses inline logging.

    32. Specified User Operator loggers and log levels.

    33. Kafka Exporter configuration. Kafka Exporter is an optional component for extracting metrics data from Kafka brokers, in particular consumer lag data.

    34. Optional configuration for Cruise Control, which is used to rebalance the Kafka cluster.

    35. Cruise Control TLS sidecar configuration. Cruise Control uses the TLS sidecar for secure communication with ZooKeeper.

  2. Create or update the resource:

    kubectl apply -f KAFKA-CONFIG-FILE

2.1.2. Configuring the Entity Operator

The Entity Operator is responsible for managing Kafka-related entities in a running Kafka cluster.

The Entity Operator comprises the:

Through Kafka resource configuration, the Cluster Operator can deploy the Entity Operator, including one or both operators, when deploying a Kafka cluster.

Note
When deployed, the Entity Operator contains the operators according to the deployment configuration.

The operators are automatically configured to manage the topics and users of the Kafka cluster.

Entity Operator configuration properties

Use the entityOperator property in Kafka.spec to configure the Entity Operator.

The entityOperator property supports several sub-properties:

  • tlsSidecar

  • topicOperator

  • userOperator

  • template

The tlsSidecar property contains the configuration of the TLS sidecar container, which is used to communicate with ZooKeeper.

The template property contains the configuration of the Entity Operator pod, such as labels, annotations, affinity, and tolerations. For more information on configuring templates, see Customizing Kubernetes resources.

The topicOperator property contains the configuration of the Topic Operator. When this option is missing, the Entity Operator is deployed without the Topic Operator.

The userOperator property contains the configuration of the User Operator. When this option is missing, the Entity Operator is deployed without the User Operator.

For more information on the properties used to configure the Entity Operator, see the EntityUserOperatorSpec schema reference.

Example of basic configuration enabling both operators
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
  zookeeper:
    # ...
  entityOperator:
    topicOperator: {}
    userOperator: {}

If an empty object ({}) is used for the topicOperator and userOperator, all properties use their default values.

When both topicOperator and userOperator properties are missing, the Entity Operator is not deployed.

Topic Operator configuration properties

Topic Operator deployment can be configured using additional options inside the topicOperator object. The following properties are supported:

watchedNamespace

The Kubernetes namespace in which the topic operator watches for KafkaTopics. Default is the namespace where the Kafka cluster is deployed.

reconciliationIntervalSeconds

The interval between periodic reconciliations in seconds. Default 120.

zookeeperSessionTimeoutSeconds

The ZooKeeper session timeout in seconds. Default 18.

topicMetadataMaxAttempts

The number of attempts at getting topic metadata from Kafka. The time between each attempt is defined as an exponential back-off. Consider increasing this value when topic creation might take more time due to the number of partitions or replicas. Default 6.

image

The image property can be used to configure the container image which will be used. For more details about configuring custom container images, see image.

resources

The resources property configures the amount of resources allocated to the Topic Operator. For more details about resource request and limit configuration, see resources.

logging

The logging property configures the logging of the Topic Operator. For more details, see logging.

Example Topic Operator configuration
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
  zookeeper:
    # ...
  entityOperator:
    # ...
    topicOperator:
      watchedNamespace: my-topic-namespace
      reconciliationIntervalSeconds: 60
    # ...
User Operator configuration properties

User Operator deployment can be configured using additional options inside the userOperator object. The following properties are supported:

watchedNamespace

The Kubernetes namespace in which the user operator watches for KafkaUsers. Default is the namespace where the Kafka cluster is deployed.

reconciliationIntervalSeconds

The interval between periodic reconciliations in seconds. Default 120.

zookeeperSessionTimeoutSeconds

The ZooKeeper session timeout in seconds. Default 18.

image

The image property can be used to configure the container image which will be used. For more details about configuring custom container images, see image.

resources

The resources property configures the amount of resources allocated to the User Operator. For more details about resource request and limit configuration, see resources.

logging

The logging property configures the logging of the User Operator. For more details, see logging.

secretPrefix

The secretPrefix property adds a prefix to the name of all Secrets created from the KafkaUser resource. For example, STRIMZI_SECRET_PREFIX=kafka- would prefix all Secret names with kafka-. So a KafkaUser named my-user would create a Secret named kafka-my-user.

Example User Operator configuration
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
  zookeeper:
    # ...
  entityOperator:
    # ...
    userOperator:
      watchedNamespace: my-user-namespace
      reconciliationIntervalSeconds: 60
    # ...

2.1.3. Kafka and ZooKeeper storage types

As stateful applications, Kafka and ZooKeeper need to store data on disk. Strimzi supports three storage types for this data:

  • Ephemeral

  • Persistent

  • JBOD storage

Note
JBOD storage is supported only for Kafka, not for ZooKeeper.

When configuring a Kafka resource, you can specify the type of storage used by the Kafka broker and its corresponding ZooKeeper node. You configure the storage type using the storage property in the following resources:

  • Kafka.spec.kafka

  • Kafka.spec.zookeeper

The storage type is configured in the type field.

Warning
The storage type cannot be changed after a Kafka cluster is deployed.
Additional resources
Data storage considerations

An efficient data storage infrastructure is essential to the optimal performance of Strimzi.

Block storage is required. File storage, such as NFS, does not work with Kafka.

Choose from one of the following options for your block storage:

Note
Strimzi does not require Kubernetes raw block volumes.
File systems

It is recommended that you configure your storage system to use the XFS file system. Strimzi is also compatible with the ext4 file system, but this might require additional configuration for best results.

Apache Kafka and ZooKeeper storage

Use separate disks for Apache Kafka and ZooKeeper.

Three types of data storage are supported:

  • Ephemeral (Recommended for development only)

  • Persistent

  • JBOD (Just a Bunch of Disks, suitable for Kafka only)

For more information, see Kafka and ZooKeeper storage.

Solid-state drives (SSDs), though not essential, can improve the performance of Kafka in large clusters where data is sent to and received from multiple topics asynchronously. SSDs are particularly effective with ZooKeeper, which requires fast, low latency data access.

Note
You do not need to provision replicated storage because Kafka and ZooKeeper both have built-in data replication.
Ephemeral storage

Ephemeral storage uses emptyDir volumes to store data. To use ephemeral storage, set the type field to ephemeral.

Important
emptyDir volumes are not persistent and the data stored in them is lost when the pod is restarted. After the new pod is started, it must recover all data from the other nodes of the cluster. Ephemeral storage is not suitable for use with single-node ZooKeeper clusters or for Kafka topics with a replication factor of 1. This configuration will cause data loss.
An example of Ephemeral storage
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
    storage:
      type: ephemeral
    # ...
  zookeeper:
    # ...
    storage:
      type: ephemeral
    # ...
Log directories

The ephemeral volume is used by the Kafka brokers as log directories mounted into the following path:

/var/lib/kafka/data/kafka-logIDX

Where IDX is the Kafka broker pod index. For example /var/lib/kafka/data/kafka-log0.

Persistent storage

Persistent storage uses Persistent Volume Claims to provision persistent volumes for storing data. Persistent Volume Claims can be used to provision volumes of many different types, depending on the Storage Class which will provision the volume. The data types which can be used with persistent volume claims include many types of SAN storage as well as Local persistent volumes.

To use persistent storage, the type has to be set to persistent-claim. Persistent storage supports additional configuration options:

id (optional)

Storage identification number. This option is mandatory for storage volumes defined in a JBOD storage declaration. Default is 0.

size (required)

Defines the size of the persistent volume claim, for example, "1000Gi".

class (optional)

The Kubernetes Storage Class to use for dynamic volume provisioning.

selector (optional)

Allows selecting a specific persistent volume to use. It contains key:value pairs representing labels for selecting such a volume.

deleteClaim (optional)

Boolean value which specifies if the Persistent Volume Claim has to be deleted when the cluster is undeployed. Default is false.

Warning
Increasing the size of persistent volumes in an existing Strimzi cluster is only supported in Kubernetes versions that support persistent volume resizing. The persistent volume to be resized must use a storage class that supports volume expansion. For other versions of Kubernetes and storage classes which do not support volume expansion, you must decide the necessary storage size before deploying the cluster. Decreasing the size of existing persistent volumes is not possible.
Example fragment of persistent storage configuration with 1000Gi size
# ...
storage:
  type: persistent-claim
  size: 1000Gi
# ...

The following example demonstrates the use of a storage class.

Example fragment of persistent storage configuration with specific Storage Class
# ...
storage:
  type: persistent-claim
  size: 1Gi
  class: my-storage-class
# ...

Finally, a selector can be used to select a specific labeled persistent volume to provide needed features such as an SSD.

Example fragment of persistent storage configuration with selector
# ...
storage:
  type: persistent-claim
  size: 1Gi
  selector:
    hdd-type: ssd
  deleteClaim: true
# ...
Storage class overrides

You can specify a different storage class for one or more Kafka brokers or ZooKeeper nodes, instead of using the default storage class. This is useful if, for example, storage classes are restricted to different availability zones or data centers. You can use the overrides field for this purpose.

In this example, the default storage class is named my-storage-class:

Example Strimzi cluster using storage class overrides
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  labels:
    app: my-cluster
  name: my-cluster
  namespace: myproject
spec:
  # ...
  kafka:
    replicas: 3
    storage:
      deleteClaim: true
      size: 100Gi
      type: persistent-claim
      class: my-storage-class
      overrides:
        - broker: 0
          class: my-storage-class-zone-1a
        - broker: 1
          class: my-storage-class-zone-1b
        - broker: 2
          class: my-storage-class-zone-1c
  # ...
  zookeeper:
    replicas: 3
    storage:
      deleteClaim: true
      size: 100Gi
      type: persistent-claim
      class: my-storage-class
      overrides:
        - broker: 0
          class: my-storage-class-zone-1a
        - broker: 1
          class: my-storage-class-zone-1b
        - broker: 2
          class: my-storage-class-zone-1c
  # ...

As a result of the configured overrides property, the volumes use the following storage classes:

  • The persistent volumes of ZooKeeper node 0 will use my-storage-class-zone-1a.

  • The persistent volumes of ZooKeeper node 1 will use my-storage-class-zone-1b.

  • The persistent volumes of ZooKeeepr node 2 will use my-storage-class-zone-1c.

  • The persistent volumes of Kafka broker 0 will use my-storage-class-zone-1a.

  • The persistent volumes of Kafka broker 1 will use my-storage-class-zone-1b.

  • The persistent volumes of Kafka broker 2 will use my-storage-class-zone-1c.

The overrides property is currently used only to override storage class configurations. Overriding other storage configuration fields is not currently supported. Other fields from the storage configuration are currently not supported.

Persistent Volume Claim naming

When persistent storage is used, it creates Persistent Volume Claims with the following names:

data-cluster-name-kafka-idx

Persistent Volume Claim for the volume used for storing data for the Kafka broker pod idx.

data-cluster-name-zookeeper-idx

Persistent Volume Claim for the volume used for storing data for the ZooKeeper node pod idx.

Log directories

The persistent volume is used by the Kafka brokers as log directories mounted into the following path:

/var/lib/kafka/data/kafka-logIDX

Where IDX is the Kafka broker pod index. For example /var/lib/kafka/data/kafka-log0.

Resizing persistent volumes

You can provision increased storage capacity by increasing the size of the persistent volumes used by an existing Strimzi cluster. Resizing persistent volumes is supported in clusters that use either a single persistent volume or multiple persistent volumes in a JBOD storage configuration.

Note
You can increase but not decrease the size of persistent volumes. Decreasing the size of persistent volumes is not currently supported in Kubernetes.
Prerequisites
  • A Kubernetes cluster with support for volume resizing.

  • The Cluster Operator is running.

  • A Kafka cluster using persistent volumes created using a storage class that supports volume expansion.

Procedure
  1. In a Kafka resource, increase the size of the persistent volume allocated to the Kafka cluster, the ZooKeeper cluster, or both.

    • To increase the volume size allocated to the Kafka cluster, edit the spec.kafka.storage property.

    • To increase the volume size allocated to the ZooKeeper cluster, edit the spec.zookeeper.storage property.

      For example, to increase the volume size from 1000Gi to 2000Gi:

      apiVersion: kafka.strimzi.io/v1beta2
      kind: Kafka
      metadata:
        name: my-cluster
      spec:
        kafka:
          # ...
          storage:
            type: persistent-claim
            size: 2000Gi
            class: my-storage-class
          # ...
        zookeeper:
          # ...
  2. Create or update the resource:

    kubectl apply -f KAFKA-CONFIG-FILE

    Kubernetes increases the capacity of the selected persistent volumes in response to a request from the Cluster Operator. When the resizing is complete, the Cluster Operator restarts all pods that use the resized persistent volumes. This happens automatically.

Additional resources

For more information about resizing persistent volumes in Kubernetes, see Resizing Persistent Volumes using Kubernetes.

JBOD storage overview

You can configure Strimzi to use JBOD, a data storage configuration of multiple disks or volumes. JBOD is one approach to providing increased data storage for Kafka brokers. It can also improve performance.

A JBOD configuration is described by one or more volumes, each of which can be either ephemeral or persistent. The rules and constraints for JBOD volume declarations are the same as those for ephemeral and persistent storage. For example, you cannot decrease the size of a persistent storage volume after it has been provisioned, or you cannot change the value of sizeLimit when type=ephemeral.

JBOD configuration

To use JBOD with Strimzi, the storage type must be set to jbod. The volumes property allows you to describe the disks that make up your JBOD storage array or configuration. The following fragment shows an example JBOD configuration:

# ...
storage:
  type: jbod
  volumes:
  - id: 0
    type: persistent-claim
    size: 100Gi
    deleteClaim: false
  - id: 1
    type: persistent-claim
    size: 100Gi
    deleteClaim: false
# ...

The ids cannot be changed once the JBOD volumes are created.

Users can add or remove volumes from the JBOD configuration.

JBOD and Persistent Volume Claims

When persistent storage is used to declare JBOD volumes, the naming scheme of the resulting Persistent Volume Claims is as follows:

data-id-cluster-name-kafka-idx

Where id is the ID of the volume used for storing data for Kafka broker pod idx.

Log directories

The JBOD volumes will be used by the Kafka brokers as log directories mounted into the following path:

/var/lib/kafka/data-id/kafka-log_idx_

Where id is the ID of the volume used for storing data for Kafka broker pod idx. For example /var/lib/kafka/data-0/kafka-log0.

Adding volumes to JBOD storage

This procedure describes how to add volumes to a Kafka cluster configured to use JBOD storage. It cannot be applied to Kafka clusters configured to use any other storage type.

Note
When adding a new volume under an id which was already used in the past and removed, you have to make sure that the previously used PersistentVolumeClaims have been deleted.
Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

  • A Kafka cluster with JBOD storage

Procedure
  1. Edit the spec.kafka.storage.volumes property in the Kafka resource. Add the new volumes to the volumes array. For example, add the new volume with id 2:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    metadata:
      name: my-cluster
    spec:
      kafka:
        # ...
        storage:
          type: jbod
          volumes:
          - id: 0
            type: persistent-claim
            size: 100Gi
            deleteClaim: false
          - id: 1
            type: persistent-claim
            size: 100Gi
            deleteClaim: false
          - id: 2
            type: persistent-claim
            size: 100Gi
            deleteClaim: false
        # ...
      zookeeper:
        # ...
  2. Create or update the resource:

    kubectl apply -f KAFKA-CONFIG-FILE
  3. Create new topics or reassign existing partitions to the new disks.

Additional resources

For more information about reassigning topics, see Partition reassignment.

Removing volumes from JBOD storage

This procedure describes how to remove volumes from Kafka cluster configured to use JBOD storage. It cannot be applied to Kafka clusters configured to use any other storage type. The JBOD storage always has to contain at least one volume.

Important
To avoid data loss, you have to move all partitions before removing the volumes.
Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

  • A Kafka cluster with JBOD storage with two or more volumes

Procedure
  1. Reassign all partitions from the disks which are you going to remove. Any data in partitions still assigned to the disks which are going to be removed might be lost.

  2. Edit the spec.kafka.storage.volumes property in the Kafka resource. Remove one or more volumes from the volumes array. For example, remove the volumes with ids 1 and 2:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    metadata:
      name: my-cluster
    spec:
      kafka:
        # ...
        storage:
          type: jbod
          volumes:
          - id: 0
            type: persistent-claim
            size: 100Gi
            deleteClaim: false
        # ...
      zookeeper:
        # ...
  3. Create or update the resource:

    kubectl apply -f KAFKA-CONFIG-FILE
Additional resources

For more information about reassigning topics, see Partition reassignment.

2.1.4. Scaling clusters

Scaling Kafka clusters
Adding brokers to a cluster

The primary way of increasing throughput for a topic is to increase the number of partitions for that topic. That works because the extra partitions allow the load of the topic to be shared between the different brokers in the cluster. However, in situations where every broker is constrained by a particular resource (typically I/O) using more partitions will not result in increased throughput. Instead, you need to add brokers to the cluster.

When you add an extra broker to the cluster, Kafka does not assign any partitions to it automatically. You must decide which partitions to move from the existing brokers to the new broker.

Once the partitions have been redistributed between all the brokers, the resource utilization of each broker should be reduced.

Removing brokers from a cluster

Because Strimzi uses StatefulSets to manage broker pods, you cannot remove any pod from the cluster. You can only remove one or more of the highest numbered pods from the cluster. For example, in a cluster of 12 brokers the pods are named cluster-name-kafka-0 up to cluster-name-kafka-11. If you decide to scale down by one broker, the cluster-name-kafka-11 will be removed.

Before you remove a broker from a cluster, ensure that it is not assigned to any partitions. You should also decide which of the remaining brokers will be responsible for each of the partitions on the broker being decommissioned. Once the broker has no assigned partitions, you can scale the cluster down safely.

Partition reassignment

The Topic Operator does not currently support reassigning replicas to different brokers, so it is necessary to connect directly to broker pods to reassign replicas to brokers.

Within a broker pod, the kafka-reassign-partitions.sh utility allows you to reassign partitions to different brokers.

It has three different modes:

--generate

Takes a set of topics and brokers and generates a reassignment JSON file which will result in the partitions of those topics being assigned to those brokers. Because this operates on whole topics, it cannot be used when you only want to reassign some partitions of some topics.

--execute

Takes a reassignment JSON file and applies it to the partitions and brokers in the cluster. Brokers that gain partitions as a result become followers of the partition leader. For a given partition, once the new broker has caught up and joined the ISR (in-sync replicas) the old broker will stop being a follower and will delete its replica.

--verify

Using the same reassignment JSON file as the --execute step, --verify checks whether all the partitions in the file have been moved to their intended brokers. If the reassignment is complete, --verify also removes any throttles that are in effect. Unless removed, throttles will continue to affect the cluster even after the reassignment has finished.

It is only possible to have one reassignment running in a cluster at any given time, and it is not possible to cancel a running reassignment. If you need to cancel a reassignment, wait for it to complete and then perform another reassignment to revert the effects of the first reassignment. The kafka-reassign-partitions.sh will print the reassignment JSON for this reversion as part of its output. Very large reassignments should be broken down into a number of smaller reassignments in case there is a need to stop in-progress reassignment.

Reassignment JSON file

The reassignment JSON file has a specific structure:

{
  "version": 1,
  "partitions": [
    <PartitionObjects>
  ]
}

Where <PartitionObjects> is a comma-separated list of objects like:

{
  "topic": <TopicName>,
  "partition": <Partition>,
  "replicas": [ <AssignedBrokerIds> ]
}
Note
Although Kafka also supports a "log_dirs" property this should not be used in Strimzi.

The following is an example reassignment JSON file that assigns partition 4 of topic topic-a to brokers 2, 4 and 7, and partition 2 of topic topic-b to brokers 1, 5 and 7:

{
  "version": 1,
  "partitions": [
    {
      "topic": "topic-a",
      "partition": 4,
      "replicas": [2,4,7]
    },
    {
      "topic": "topic-b",
      "partition": 2,
      "replicas": [1,5,7]
    }
  ]
}

Partitions not included in the JSON are not changed.

Reassigning partitions between JBOD volumes

When using JBOD storage in your Kafka cluster, you can choose to reassign the partitions between specific volumes and their log directories (each volume has a single log directory). To reassign a partition to a specific volume, add the log_dirs option to <PartitionObjects> in the reassignment JSON file.

{
  "topic": <TopicName>,
  "partition": <Partition>,
  "replicas": [ <AssignedBrokerIds> ],
  "log_dirs": [ <AssignedLogDirs> ]
}

The log_dirs object should contain the same number of log directories as the number of replicas specified in the replicas object. The value should be either an absolute path to the log directory, or the any keyword.

For example:

{
      "topic": "topic-a",
      "partition": 4,
      "replicas": [2,4,7].
      "log_dirs": [ "/var/lib/kafka/data-0/kafka-log2", "/var/lib/kafka/data-0/kafka-log4", "/var/lib/kafka/data-0/kafka-log7" ]
}
Generating reassignment JSON files

This procedure describes how to generate a reassignment JSON file that reassigns all the partitions for a given set of topics using the kafka-reassign-partitions.sh tool.

Prerequisites
  • A running Cluster Operator

  • A Kafka resource

  • A set of topics to reassign the partitions of

Procedure
  1. Prepare a JSON file named topics.json that lists the topics to move. It must have the following structure:

    {
      "version": 1,
      "topics": [
        <TopicObjects>
      ]
    }

    where <TopicObjects> is a comma-separated list of objects like:

    {
      "topic": <TopicName>
    }

    For example if you want to reassign all the partitions of topic-a and topic-b, you would need to prepare a topics.json file like this:

    {
      "version": 1,
      "topics": [
        { "topic": "topic-a"},
        { "topic": "topic-b"}
      ]
    }
  2. Copy the topics.json file to one of the broker pods:

    cat topics.json | kubectl exec -c kafka <BrokerPod> -i -- \
      /bin/bash -c \
      'cat > /tmp/topics.json'
  3. Use the kafka-reassign-partitions.sh command to generate the reassignment JSON.

    kubectl exec <BrokerPod> -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --topics-to-move-json-file /tmp/topics.json \
      --broker-list <BrokerList> \
      --generate

    For example, to move all the partitions of topic-a and topic-b to brokers 4 and 7

    kubectl exec <BrokerPod> -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --topics-to-move-json-file /tmp/topics.json \
      --broker-list 4,7 \
      --generate
Creating reassignment JSON files manually

You can manually create the reassignment JSON file if you want to move specific partitions.

Reassignment throttles

Partition reassignment can be a slow process because it involves transferring large amounts of data between brokers. To avoid a detrimental impact on clients, you can throttle the reassignment process. This might cause the reassignment to take longer to complete.

  • If the throttle is too low then the newly assigned brokers will not be able to keep up with records being published and the reassignment will never complete.

  • If the throttle is too high then clients will be impacted.

For example, for producers, this could manifest as higher than normal latency waiting for acknowledgement. For consumers, this could manifest as a drop in throughput caused by higher latency between polls.

Scaling up a Kafka cluster

This procedure describes how to increase the number of brokers in a Kafka cluster.

Prerequisites
  • An existing Kafka cluster.

  • A reassignment JSON file named reassignment.json that describes how partitions should be reassigned to brokers in the enlarged cluster.

Procedure
  1. Add as many new brokers as you need by increasing the Kafka.spec.kafka.replicas configuration option.

  2. Verify that the new broker pods have started.

  3. Copy the reassignment.json file to the broker pod on which you will later execute the commands:

    cat reassignment.json | \
      kubectl exec broker-pod -c kafka -i -- /bin/bash -c \
      'cat > /tmp/reassignment.json'

    For example:

    cat reassignment.json | \
      kubectl exec my-cluster-kafka-0 -c kafka -i -- /bin/bash -c \
      'cat > /tmp/reassignment.json'
  4. Execute the partition reassignment using the kafka-reassign-partitions.sh command line tool from the same broker pod.

    kubectl exec broker-pod -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --execute

    If you are going to throttle replication you can also pass the --throttle option with an inter-broker throttled rate in bytes per second. For example:

    kubectl exec my-cluster-kafka-0 -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --throttle 5000000 \
      --execute

    This command will print out two reassignment JSON objects. The first records the current assignment for the partitions being moved. You should save this to a local file (not a file in the pod) in case you need to revert the reassignment later on. The second JSON object is the target reassignment you have passed in your reassignment JSON file.

  5. If you need to change the throttle during reassignment you can use the same command line with a different throttled rate. For example:

    kubectl exec my-cluster-kafka-0 -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --throttle 10000000 \
      --execute
  6. Periodically verify whether the reassignment has completed using the kafka-reassign-partitions.sh command line tool from any of the broker pods. This is the same command as the previous step but with the --verify option instead of the --execute option.

    kubectl exec broker-pod -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --verify

    For example,

    kubectl exec my-cluster-kafka-0 -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --verify
  7. The reassignment has finished when the --verify command reports each of the partitions being moved as completed successfully. This final --verify will also have the effect of removing any reassignment throttles. You can now delete the revert file if you saved the JSON for reverting the assignment to their original brokers.

Scaling down a Kafka cluster

This procedure describes how to decrease the number of brokers in a Kafka cluster.

Prerequisites
  • An existing Kafka cluster.

  • A reassignment JSON file named reassignment.json describing how partitions should be reassigned to brokers in the cluster once the broker(s) in the highest numbered Pod(s) have been removed.

Procedure
  1. Copy the reassignment.json file to the broker pod on which you will later execute the commands:

    cat reassignment.json | \
      kubectl exec broker-pod -c kafka -i -- /bin/bash -c \
      'cat > /tmp/reassignment.json'

    For example:

    cat reassignment.json | \
      kubectl exec my-cluster-kafka-0 -c kafka -i -- /bin/bash -c \
      'cat > /tmp/reassignment.json'
  2. Execute the partition reassignment using the kafka-reassign-partitions.sh command line tool from the same broker pod.

    kubectl exec broker-pod -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --execute

    If you are going to throttle replication you can also pass the --throttle option with an inter-broker throttled rate in bytes per second. For example:

    kubectl exec my-cluster-kafka-0 -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --throttle 5000000 \
      --execute

    This command will print out two reassignment JSON objects. The first records the current assignment for the partitions being moved. You should save this to a local file (not a file in the pod) in case you need to revert the reassignment later on. The second JSON object is the target reassignment you have passed in your reassignment JSON file.

  3. If you need to change the throttle during reassignment you can use the same command line with a different throttled rate. For example:

    kubectl exec my-cluster-kafka-0 -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --throttle 10000000 \
      --execute
  4. Periodically verify whether the reassignment has completed using the kafka-reassign-partitions.sh command line tool from any of the broker pods. This is the same command as the previous step but with the --verify option instead of the --execute option.

    kubectl exec broker-pod -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --verify

    For example,

    kubectl exec my-cluster-kafka-0 -c kafka -it -- \
      bin/kafka-reassign-partitions.sh --bootstrap-server localhost:9092 \
      --reassignment-json-file /tmp/reassignment.json \
      --verify
  5. The reassignment has finished when the --verify command reports each of the partitions being moved as completed successfully. This final --verify will also have the effect of removing any reassignment throttles. You can now delete the revert file if you saved the JSON for reverting the assignment to their original brokers.

  6. Once all the partition reassignments have finished, the broker(s) being removed should not have responsibility for any of the partitions in the cluster. You can verify this by checking that the broker’s data log directory does not contain any live partition logs. If the log directory on the broker contains a directory that does not match the extended regular expression [a-zA-Z0-9.-]+\.[a-z0-9]+-delete$ then the broker still has live partitions and it should not be stopped.

    You can check this by executing the command:

    kubectl exec my-cluster-kafka-0 -c kafka -it -- \
      /bin/bash -c \
      "ls -l /var/lib/kafka/kafka-log_<N>_ | grep -E '^d' | grep -vE '[a-zA-Z0-9.-]+\.[a-z0-9]+-delete$'"

    where N is the number of the Pod(s) being deleted.

    If the above command prints any output then the broker still has live partitions. In this case, either the reassignment has not finished, or the reassignment JSON file was incorrect.

  7. Once you have confirmed that the broker has no live partitions you can edit the Kafka.spec.kafka.replicas of your Kafka resource, which will scale down the StatefulSet, deleting the highest numbered broker Pod(s).

2.1.5. Retrieving JMX metrics with JmxTrans

JmxTrans is a tool for retrieving JMX metrics data from Java processes and pushing that data, in various formats, to remote sinks inside or outside the cluster. JmxTrans can communicate with a secure JMX port.

Strimzi supports using JmxTrans to read JMX metrics from Kafka brokers.

JmxTrans reads JMX metrics data from secure or insecure Kafka brokers and pushes the data to remote sinks in various data formats. For example, JmxTrans can obtain JMX metrics about the request rate of each Kafka broker’s network and then push the data to a Logstash database outside the Kubernetes cluster.

Configuring a JmxTrans deployment
Prerequisites
  • A running Kubernetes cluster

You can configure a JmxTrans deployment by using the Kafka.spec.jmxTrans property. A JmxTrans deployment can read from a secure or insecure Kafka broker. To configure a JmxTrans deployment, define the following properties:

  • Kafka.spec.jmxTrans.outputDefinitions

  • Kafka.spec.jmxTrans.kafkaQueries

For more information on these properties, see the JmxTransSpec schema reference.

Configuring JmxTrans output definitions

Output definitions specify where JMX metrics are pushed to, and in which data format. For information about supported data formats, see Data formats. How many seconds JmxTrans agent waits for before pushing new data can be configured through the flushDelay property. The host and port properties define the target host address and target port the data is pushed to. The name property is a required property that is referenced by the Kafka.spec.jmxTrans.kafkaQueries property.

Here is an example configuration pushing JMX data in the Graphite format every 5 seconds to a Logstash database on http://myLogstash:9999, and another pushing to standardOut (standard output):

apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  jmxTrans:
    outputDefinitions:
      - outputType: "com.googlecode.jmxtrans.model.output.GraphiteWriter"
        host: "http://myLogstash"
        port: 9999
        flushDelay: 5
        name: "logstash"
      - outputType: "com.googlecode.jmxtrans.model.output.StdOutWriter"
        name: "standardOut"
        # ...
    # ...
  zookeeper:
    # ...
Configuring JmxTrans queries

JmxTrans queries specify what JMX metrics are read from the Kafka brokers. Currently JmxTrans queries can only be sent to the Kafka Brokers. Configure the targetMBean property to specify which target MBean on the Kafka broker is addressed. Configuring the attributes property specifies which MBean attribute is read as JMX metrics from the target MBean. JmxTrans supports wildcards to read from target MBeans, and filter by specifying the typenames. The outputs property defines where the metrics are pushed to by specifying the name of the output definitions.

The following JmxTrans deployment reads from all MBeans that match the pattern kafka.server:type=BrokerTopicMetrics,name=* and have name in the target MBean’s name. From those Mbeans, it obtains JMX metrics about the Count attribute and pushes the metrics to standard output as defined by outputs.

apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  # ...
  jmxTrans:
    kafkaQueries:
      - targetMBean: "kafka.server:type=BrokerTopicMetrics,*"
        typeNames: ["name"]
        attributes:  ["Count"]
        outputs: ["standardOut"]
  zookeeper:
    # ...
Additional resources

For more information about JmxTrans, see the JmxTrans github.

2.1.6. Maintenance time windows for rolling updates

Maintenance time windows allow you to schedule certain rolling updates of your Kafka and ZooKeeper clusters to start at a convenient time.

Maintenance time windows overview

In most cases, the Cluster Operator only updates your Kafka or ZooKeeper clusters in response to changes to the corresponding Kafka resource. This enables you to plan when to apply changes to a Kafka resource to minimize the impact on Kafka client applications.

However, some updates to your Kafka and ZooKeeper clusters can happen without any corresponding change to the Kafka resource. For example, the Cluster Operator will need to perform a rolling restart if a CA (Certificate Authority) certificate that it manages is close to expiry.

While a rolling restart of the pods should not affect availability of the service (assuming correct broker and topic configurations), it could affect performance of the Kafka client applications. Maintenance time windows allow you to schedule such spontaneous rolling updates of your Kafka and ZooKeeper clusters to start at a convenient time. If maintenance time windows are not configured for a cluster then it is possible that such spontaneous rolling updates will happen at an inconvenient time, such as during a predictable period of high load.

Maintenance time window definition

You configure maintenance time windows by entering an array of strings in the Kafka.spec.maintenanceTimeWindows property. Each string is a cron expression interpreted as being in UTC (Coordinated Universal Time, which for practical purposes is the same as Greenwich Mean Time).

The following example configures a single maintenance time window that starts at midnight and ends at 01:59am (UTC), on Sundays, Mondays, Tuesdays, Wednesdays, and Thursdays:

# ...
maintenanceTimeWindows:
  - "* * 0-1 ? * SUN,MON,TUE,WED,THU *"
# ...

In practice, maintenance windows should be set in conjunction with the Kafka.spec.clusterCa.renewalDays and Kafka.spec.clientsCa.renewalDays properties of the Kafka resource, to ensure that the necessary CA certificate renewal can be completed in the configured maintenance time windows.

Note
Strimzi does not schedule maintenance operations exactly according to the given windows. Instead, for each reconciliation, it checks whether a maintenance window is currently "open". This means that the start of maintenance operations within a given time window can be delayed by up to the Cluster Operator reconciliation interval. Maintenance time windows must therefore be at least this long.
Additional resources
Configuring a maintenance time window

You can configure a maintenance time window for rolling updates triggered by supported processes.

Prerequisites
  • A Kubernetes cluster.

  • The Cluster Operator is running.

Procedure
  1. Add or edit the maintenanceTimeWindows property in the Kafka resource. For example to allow maintenance between 0800 and 1059 and between 1400 and 1559 you would set the maintenanceTimeWindows as shown below:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    metadata:
      name: my-cluster
    spec:
      kafka:
        # ...
      zookeeper:
        # ...
      maintenanceTimeWindows:
        - "* * 8-10 * * ?"
        - "* * 14-15 * * ?"
  2. Create or update the resource:

    kubectl apply -f KAFKA-CONFIG-FILE
Additional resources

Performing rolling updates:

2.1.7. Connecting to ZooKeeper from a terminal

Most Kafka CLI tools can connect directly to Kafka, so under normal circumstances you should not need to connect to ZooKeeper. ZooKeeper services are secured with encryption and authentication and are not intended to be used by external applications that are not part of Strimzi.

However, if you want to use Kafka CLI tools that require a connection to ZooKeeper, you can use a terminal inside a ZooKeeper container and connect to localhost:12181 as the ZooKeeper address.

Prerequisites
  • A Kubernetes cluster is available.

  • A Kafka cluster is running.

  • The Cluster Operator is running.

Procedure
  1. Open the terminal using the Kubernetes console or run the exec command from your CLI.

    For example:

    kubectl exec -ti my-cluster-zookeeper-0 -- bin/kafka-topics.sh --list --zookeeper localhost:12181

    Be sure to use localhost:12181.

    You can now run Kafka commands to ZooKeeper.

2.1.8. Deleting Kafka nodes manually

This procedure describes how to delete an existing Kafka node by using a Kubernetes annotation. Deleting a Kafka node consists of deleting both the Pod on which the Kafka broker is running and the related PersistentVolumeClaim (if the cluster was deployed with persistent storage). After deletion, the Pod and its related PersistentVolumeClaim are recreated automatically.

Warning
Deleting a PersistentVolumeClaim can cause permanent data loss. The following procedure should only be performed if you have encountered storage issues.
Prerequisites

See the Deploying and Upgrading Strimzi guide for instructions on running a:

Procedure
  1. Find the name of the Pod that you want to delete.

    For example, if the cluster is named cluster-name, the pods are named cluster-name-kafka-index, where index starts at zero and ends at the total number of replicas.

  2. Annotate the Pod resource in Kubernetes.

    Use kubectl annotate:

    kubectl annotate pod cluster-name-kafka-index strimzi.io/delete-pod-and-pvc=true
  3. Wait for the next reconciliation, when the annotated pod with the underlying persistent volume claim will be deleted and then recreated.

2.1.9. Deleting ZooKeeper nodes manually

This procedure describes how to delete an existing ZooKeeper node by using a Kubernetes annotation. Deleting a ZooKeeper node consists of deleting both the Pod on which ZooKeeper is running and the related PersistentVolumeClaim (if the cluster was deployed with persistent storage). After deletion, the Pod and its related PersistentVolumeClaim are recreated automatically.

Warning
Deleting a PersistentVolumeClaim can cause permanent data loss. The following procedure should only be performed if you have encountered storage issues.
Prerequisites

See the Deploying and Upgrading Strimzi guide for instructions on running a:

Procedure
  1. Find the name of the Pod that you want to delete.

    For example, if the cluster is named cluster-name, the pods are named cluster-name-zookeeper-index, where index starts at zero and ends at the total number of replicas.

  2. Annotate the Pod resource in Kubernetes.

    Use kubectl annotate:

    kubectl annotate pod cluster-name-zookeeper-index strimzi.io/delete-pod-and-pvc=true
  3. Wait for the next reconciliation, when the annotated pod with the underlying persistent volume claim will be deleted and then recreated.

2.1.10. List of Kafka cluster resources

The following resources are created by the Cluster Operator in the Kubernetes cluster:

Shared resources
cluster-name-cluster-ca

Secret with the Cluster CA private key used to encrypt the cluster communication.

cluster-name-cluster-ca-cert

Secret with the Cluster CA public key. This key can be used to verify the identity of the Kafka brokers.

cluster-name-clients-ca

Secret with the Clients CA private key used to sign user certificates

cluster-name-clients-ca-cert

Secret with the Clients CA public key. This key can be used to verify the identity of the Kafka users.

cluster-name-cluster-operator-certs

Secret with Cluster operators keys for communication with Kafka and ZooKeeper.

Zookeeper nodes
cluster-name-zookeeper

StatefulSet which is in charge of managing the ZooKeeper node pods.

cluster-name-zookeeper-idx

Pods created by the Zookeeper StatefulSet.

cluster-name-zookeeper-nodes

Headless Service needed to have DNS resolve the ZooKeeper pods IP addresses directly.

cluster-name-zookeeper-client

Service used by Kafka brokers to connect to ZooKeeper nodes as clients.

cluster-name-zookeeper-config

ConfigMap that contains the ZooKeeper ancillary configuration, and is mounted as a volume by the ZooKeeper node pods.

cluster-name-zookeeper-nodes

Secret with ZooKeeper node keys.

cluster-name-zookeeper

Service account used by the Zookeeper nodes.

cluster-name-zookeeper

Pod Disruption Budget configured for the ZooKeeper nodes.

cluster-name-network-policy-zookeeper

Network policy managing access to the ZooKeeper services.

data-cluster-name-zookeeper-idx

Persistent Volume Claim for the volume used for storing data for the ZooKeeper node pod idx. This resource will be created only if persistent storage is selected for provisioning persistent volumes to store data.

Kafka brokers
cluster-name-kafka

StatefulSet which is in charge of managing the Kafka broker pods.

cluster-name-kafka-idx

Pods created by the Kafka StatefulSet.

cluster-name-kafka-brokers

Service needed to have DNS resolve the Kafka broker pods IP addresses directly.

cluster-name-kafka-bootstrap

Service can be used as bootstrap servers for Kafka clients connecting from within the Kubernetes cluster.

cluster-name-kafka-external-bootstrap

Bootstrap service for clients connecting from outside the Kubernetes cluster. This resource is created only when an external listener is enabled. The old service name will be used for backwards compatibility when the listener name is external and port is 9094.

cluster-name-kafka-pod-id

Service used to route traffic from outside the Kubernetes cluster to individual pods. This resource is created only when an external listener is enabled. The old service name will be used for backwards compatibility when the listener name is external and port is 9094.

cluster-name-kafka-external-bootstrap

Bootstrap route for clients connecting from outside the Kubernetes cluster. This resource is created only when an external listener is enabled and set to type route. The old route name will be used for backwards compatibility when the listener name is external and port is 9094.

cluster-name-kafka-pod-id

Route for traffic from outside the Kubernetes cluster to individual pods. This resource is created only when an external listener is enabled and set to type route. The old route name will be used for backwards compatibility when the listener name is external and port is 9094.

cluster-name-kafka-listener-name-bootstrap

Bootstrap service for clients connecting from outside the Kubernetes cluster. This resource is created only when an external listener is enabled. The new service name will be used for all other external listeners.

cluster-name-kafka-listener-name-pod-id

Service used to route traffic from outside the Kubernetes cluster to individual pods. This resource is created only when an external listener is enabled. The new service name will be used for all other external listeners.

cluster-name-kafka-listener-name-bootstrap

Bootstrap route for clients connecting from outside the Kubernetes cluster. This resource is created only when an external listener is enabled and set to type route. The new route name will be used for all other external listeners.

cluster-name-kafka-listener-name-pod-id

Route for traffic from outside the Kubernetes cluster to individual pods. This resource is created only when an external listener is enabled and set to type route. The new route name will be used for all other external listeners.

cluster-name-kafka-config

ConfigMap which contains the Kafka ancillary configuration and is mounted as a volume by the Kafka broker pods.

cluster-name-kafka-brokers

Secret with Kafka broker keys.

cluster-name-kafka

Service account used by the Kafka brokers.

cluster-name-kafka

Pod Disruption Budget configured for the Kafka brokers.

cluster-name-network-policy-kafka

Network policy managing access to the Kafka services.

strimzi-namespace-name-cluster-name-kafka-init

Cluster role binding used by the Kafka brokers.

cluster-name-jmx

Secret with JMX username and password used to secure the Kafka broker port. This resource is created only when JMX is enabled in Kafka.

data-cluster-name-kafka-idx

Persistent Volume Claim for the volume used for storing data for the Kafka broker pod idx. This resource is created only if persistent storage is selected for provisioning persistent volumes to store data.

data-id-cluster-name-kafka-idx

Persistent Volume Claim for the volume id used for storing data for the Kafka broker pod idx. This resource is created only if persistent storage is selected for JBOD volumes when provisioning persistent volumes to store data.

Entity Operator

These resources are only created if the Entity Operator is deployed using the Cluster Operator.

cluster-name-entity-operator

Deployment with Topic and User Operators.

cluster-name-entity-operator-random-string

Pod created by the Entity Operator deployment.

cluster-name-entity-topic-operator-config

ConfigMap with ancillary configuration for Topic Operators.

cluster-name-entity-user-operator-config

ConfigMap with ancillary configuration for User Operators.

cluster-name-entity-operator-certs

Secret with Entity Operator keys for communication with Kafka and ZooKeeper.

cluster-name-entity-operator

Service account used by the Entity Operator.

strimzi-cluster-name-entity-topic-operator

Role binding used by the Entity Topic Operator.

strimzi-cluster-name-entity-user-operator

Role binding used by the Entity User Operator.

Kafka Exporter

These resources are only created if the Kafka Exporter is deployed using the Cluster Operator.

cluster-name-kafka-exporter

Deployment with Kafka Exporter.

cluster-name-kafka-exporter-random-string

Pod created by the Kafka Exporter deployment.

cluster-name-kafka-exporter

Service used to collect consumer lag metrics.

cluster-name-kafka-exporter

Service account used by the Kafka Exporter.

Cruise Control

These resources are only created if Cruise Control was deployed using the Cluster Operator.

cluster-name-cruise-control

Deployment with Cruise Control.

cluster-name-cruise-control-random-string

Pod created by the Cruise Control deployment.

cluster-name-cruise-control-config

ConfigMap that contains the Cruise Control ancillary configuration, and is mounted as a volume by the Cruise Control pods.

cluster-name-cruise-control-certs

Secret with Cruise Control keys for communication with Kafka and ZooKeeper.

cluster-name-cruise-control

Service used to communicate with Cruise Control.

cluster-name-cruise-control

Service account used by Cruise Control.

cluster-name-network-policy-cruise-control

Network policy managing access to the Cruise Control service.

JMXTrans

These resources are only created if JMXTrans is deployed using the Cluster Operator.

cluster-name-jmxtrans

Deployment with JMXTrans.

cluster-name-jmxtrans-random-string

Pod created by the JMXTrans deployment.

cluster-name-jmxtrans-config

ConfigMap that contains the JMXTrans ancillary configuration, and is mounted as a volume by the JMXTrans pods.

cluster-name-jmxtrans

Service account used by JMXTrans.

2.2. Kafka Connect/S2I cluster configuration

This section describes how to configure a Kafka Connect or Kafka Connect with Source-to-Image (S2I) deployment in your Strimzi cluster.

Kafka Connect is an integration toolkit for streaming data between Kafka brokers and other systems using connector plugins. Kafka Connect provides a framework for integrating Kafka with an external data source or target, such as a database, for import or export of data using connectors. Connectors are plugins that provide the connection configuration needed.

If you are using Kafka Connect, you configure either the KafkaConnect or the KafkaConnectS2I resource. Use the KafkaConnectS2I resource if you are using the Source-to-Image (S2I) framework to deploy Kafka Connect.

Important
With the introduction of build configuration to the KafkaConnect resource, Strimzi can now automatically build a container image with the connector plugins you require for your data connections. As a result, support for Kafka Connect with Source-to-Image (S2I) is deprecated and will be removed after Strimzi 0.24.0. To prepare for this change, you can migrate Kafka Connect S2I instances to Kafka Connect instances.

2.2.1. Configuring Kafka Connect

Use Kafka Connect to set up external data connections to your Kafka cluster.

Use the properties of the KafkaConnect or KafkaConnectS2I resource to configure your Kafka Connect deployment. The example shown in this procedure is for the KafkaConnect resource, but the properties are the same for the KafkaConnectS2I resource.

Kafka connector configuration

KafkaConnector resources allow you to create and manage connector instances for Kafka Connect in a Kubernetes-native way.

In your Kafka Connect configuration, you enable KafkaConnectors for a Kafka Connect cluster by adding the strimzi.io/use-connector-resources annotation. You can also add a build configuration so that Strimzi automatically builds a container image with the connector plugins you require for your data connections. External configuration for Kafka Connect connectors is specified through the externalConfiguration property.

To manage connectors, you can use the Kafka Connect REST API, or use KafkaConnector custom resources. KafkaConnector resources must be deployed to the same namespace as the Kafka Connect cluster they link to. For more information on using these methods to create, reconfigure, or delete connectors, see Creating and managing connectors in the Deploying and Upgrading Strimzi guide.

Connector configuration is passed to Kafka Connect as part of an HTTP request and stored within Kafka itself. ConfigMaps and Secrets are standard Kubernetes resources used for storing configurations and confidential data. You can use ConfigMaps and Secrets to configure certain elements of a connector. You can then reference the configuration values in HTTP REST commands, which keeps the configuration separate and more secure, if needed. This method applies especially to confidential data, such as usernames, passwords, or certificates.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

See the Deploying and Upgrading Strimzi guide for instructions on running a:

Procedure
  1. Edit the spec properties for the KafkaConnect or KafkaConnectS2I resource.

    The properties you can configure are shown in this example configuration:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaConnect (1)
    metadata:
      name: my-connect-cluster
      annotations:
        strimzi.io/use-connector-resources: "true" (2)
    spec:
      replicas: 3 (3)
      authentication: (4)
        type: tls
        certificateAndKey:
          certificate: source.crt
          key: source.key
          secretName: my-user-source
      bootstrapServers: my-cluster-kafka-bootstrap:9092 (5)
      tls: (6)
        trustedCertificates:
          - secretName: my-cluster-cluster-cert
            certificate: ca.crt
          - secretName: my-cluster-cluster-cert
            certificate: ca2.crt
      config: (7)
        group.id: my-connect-cluster
        offset.storage.topic: my-connect-cluster-offsets
        config.storage.topic: my-connect-cluster-configs
        status.storage.topic: my-connect-cluster-status
        key.converter: org.apache.kafka.connect.json.JsonConverter
        value.converter: org.apache.kafka.connect.json.JsonConverter
        key.converter.schemas.enable: true
        value.converter.schemas.enable: true
        config.storage.replication.factor: 3
        offset.storage.replication.factor: 3
        status.storage.replication.factor: 3
      build: (8)
        output: (9)
          type: docker
          image: my-registry.io/my-org/my-connect-cluster:latest
          pushSecret: my-registry-credentials
        plugins: (10)
          - name: debezium-postgres-connector
            artifacts:
              - type: tgz
                url: https://repo1.maven.org/maven2/io/debezium/debezium-connector-postgres/1.3.1.Final/debezium-connector-postgres-1.3.1.Final-plugin.tar.gz
                sha512sum: 962a12151bdf9a5a30627eebac739955a4fd95a08d373b86bdcea2b4d0c27dd6e1edd5cb548045e115e33a9e69b1b2a352bee24df035a0447cb820077af00c03
          - name: camel-telegram
            artifacts:
              - type: tgz
                url: https://repo.maven.apache.org/maven2/org/apache/camel/kafkaconnector/camel-telegram-kafka-connector/0.7.0/camel-telegram-kafka-connector-0.7.0-package.tar.gz
                sha512sum: a9b1ac63e3284bea7836d7d24d84208c49cdf5600070e6bd1535de654f6920b74ad950d51733e8020bf4187870699819f54ef5859c7846ee4081507f48873479
      externalConfiguration: (11)
        env:
          - name: AWS_ACCESS_KEY_ID
            valueFrom:
              secretKeyRef:
                name: aws-creds
                key: awsAccessKey
          - name: AWS_SECRET_ACCESS_KEY
            valueFrom:
              secretKeyRef:
                name: aws-creds
                key: awsSecretAccessKey
      resources: (12)
        requests:
          cpu: "1"
          memory: 2Gi
        limits:
          cpu: "2"
          memory: 2Gi
      logging: (13)
        type: inline
        loggers:
          log4j.rootLogger: "INFO"
      readinessProbe: (14)
        initialDelaySeconds: 15
        timeoutSeconds: 5
      livenessProbe:
        initialDelaySeconds: 15
        timeoutSeconds: 5
      metricsConfig: (15)
        type: jmxPrometheusExporter
        valueFrom:
          configMapKeyRef:
            name: my-config-map
            key: my-key
      jvmOptions: (16)
        "-Xmx": "1g"
        "-Xms": "1g"
      image: my-org/my-image:latest (17)
      rack:
        topologyKey: topology.kubernetes.io/zone (18)
      template: (19)
        pod:
          affinity:
            podAntiAffinity:
              requiredDuringSchedulingIgnoredDuringExecution:
                - labelSelector:
                    matchExpressions:
                      - key: application
                        operator: In
                        values:
                          - postgresql
                          - mongodb
                  topologyKey: "kubernetes.io/hostname"
        connectContainer: (20)
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
    1. Use KafkaConnect or KafkaConnectS2I, as required.

    2. Enables KafkaConnectors for the Kafka Connect cluster.

    3. The number of replica nodes.

    4. Authentication for the Kafka Connect cluster, using the TLS mechanism, as shown here, using OAuth bearer tokens, or a SASL-based SCRAM-SHA-512 or PLAIN mechanism. By default, Kafka Connect connects to Kafka brokers using a plain text connection.

    5. Bootstrap server for connection to the Kafka Connect cluster.

    6. TLS encryption with key names under which TLS certificates are stored in X.509 format for the cluster. If certificates are stored in the same secret, it can be listed multiple times.

    7. Kafka Connect configuration of workers (not connectors). Standard Apache Kafka configuration may be provided, restricted to those properties not managed directly by Strimzi.

    8. Build configuration properties for building a container image with connector plugins automatically.

    9. (Required) Configuration of the container registry where new images are pushed.

    10. (Required) List of connector plugins and their artifacts to add to the new container image. Each plugin must be configured with at least one artifact.

    11. External configuration for Kafka connectors using environment variables, as shown here, or volumes. You can also use the Kubernetes Configuration Provider to load configuration values from external sources.

    12. Requests for reservation of supported resources, currently cpu and memory, and limits to specify the maximum resources that can be consumed.

    13. Specified Kafka Connect loggers and log levels added directly (inline) or indirectly (external) through a ConfigMap. A custom ConfigMap must be placed under the log4j.properties or log4j2.properties key. For the Kafka Connect log4j.rootLogger logger, you can set the log level to INFO, ERROR, WARN, TRACE, DEBUG, FATAL or OFF.

    14. Healthchecks to know when to restart a container (liveness) and when a container can accept traffic (readiness).

    15. Prometheus metrics, which are enabled by referencing a ConfigMap containing configuration for the Prometheus JMX exporter in this example. You can enable metrics without further configuration using a reference to a ConfigMap containing an empty file under metricsConfig.valueFrom.configMapKeyRef.key.

    16. JVM configuration options to optimize performance for the Virtual Machine (VM) running Kafka Connect.

    17. ADVANCED OPTION: Container image configuration, which is recommended only in special situations.

    18. Rack awareness is configured to spread replicas across different racks. A topologykey must match the label of a cluster node.

    19. Template customization. Here a pod is scheduled with anti-affinity, so the pod is not scheduled on nodes with the same hostname.

    20. Environment variables are also set for distributed tracing using Jaeger.

  2. Create or update the resource:

    kubectl apply -f KAFKA-CONNECT-CONFIG-FILE
  3. If authorization is enabled for Kafka Connect, configure Kafka Connect users to enable access to the Kafka Connect consumer group and topics.

2.2.2. Kafka Connect configuration for multiple instances

If you are running multiple instances of Kafka Connect, you have to change the default configuration of the following config properties:

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaConnect
metadata:
  name: my-connect
spec:
  # ...
  config:
    group.id: connect-cluster (1)
    offset.storage.topic: connect-cluster-offsets (2)
    config.storage.topic: connect-cluster-configs (3)
    status.storage.topic: connect-cluster-status  (4)
    # ...
# ...
  1. Kafka Connect cluster group that the instance belongs to.

  2. Kafka topic that stores connector offsets.

  3. Kafka topic that stores connector and task status configurations.

  4. Kafka topic that stores connector and task status updates.

Note
Values for the three topics must be the same for all Kafka Connect instances with the same group.id.

Unless you change the default settings, each Kafka Connect instance connecting to the same Kafka cluster is deployed with the same values. What happens, in effect, is all instances are coupled to run in a cluster and use the same topics.

If multiple Kafka Connect clusters try to use the same topics, Kafka Connect will not work as expected and generate errors.

If you wish to run multiple Kafka Connect instances, change the values of these properties for each instance.

2.2.3. Configuring Kafka Connect user authorization

This procedure describes how to authorize user access to Kafka Connect.

When any type of authorization is being used in Kafka, a Kafka Connect user requires read/write access rights to the consumer group and the internal topics of Kafka Connect.

The properties for the consumer group and internal topics are automatically configured by Strimzi, or they can be specified explicitly in the spec of the KafkaConnect or KafkaConnectS2I resource.

Example configuration properties in the KafkaConnect resource
apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaConnect
metadata:
  name: my-connect
spec:
  # ...
  config:
    group.id: my-connect-cluster (1)
    offset.storage.topic: my-connect-cluster-offsets (2)
    config.storage.topic: my-connect-cluster-configs (3)
    status.storage.topic: my-connect-cluster-status (4)
    # ...
  # ...
  1. Kafka Connect cluster group that the instance belongs to.

  2. Kafka topic that stores connector offsets.

  3. Kafka topic that stores connector and task status configurations.

  4. Kafka topic that stores connector and task status updates.

This procedure shows how access is provided when simple authorization is being used.

Simple authorization uses ACL rules, handled by the Kafka AclAuthorizer plugin, to provide the right level of access. For more information on configuring a KafkaUser resource to use simple authorization, see the AclRule schema reference.

Note
The default values for the consumer group and topics will differ when running multiple instances.
Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Edit the authorization property in the KafkaUser resource to provide access rights to the user.

    In the following example, access rights are configured for the Kafka Connect topics and consumer group using literal name values:

    Property Name

    offset.storage.topic

    connect-cluster-offsets

    status.storage.topic

    connect-cluster-status

    config.storage.topic

    connect-cluster-configs

    group

    connect-cluster

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaUser
    metadata:
      name: my-user
      labels:
        strimzi.io/cluster: my-cluster
    spec:
      # ...
      authorization:
        type: simple
        acls:
          # access to offset.storage.topic
          - resource:
              type: topic
              name: connect-cluster-offsets
              patternType: literal
            operation: Write
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-offsets
              patternType: literal
            operation: Create
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-offsets
              patternType: literal
            operation: Describe
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-offsets
              patternType: literal
            operation: Read
            host: "*"
          # access to status.storage.topic
          - resource:
              type: topic
              name: connect-cluster-status
              patternType: literal
            operation: Write
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-status
              patternType: literal
            operation: Create
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-status
              patternType: literal
            operation: Describe
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-status
              patternType: literal
            operation: Read
            host: "*"
          # access to config.storage.topic
          - resource:
              type: topic
              name: connect-cluster-configs
              patternType: literal
            operation: Write
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-configs
              patternType: literal
            operation: Create
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-configs
              patternType: literal
            operation: Describe
            host: "*"
          - resource:
              type: topic
              name: connect-cluster-configs
              patternType: literal
            operation: Read
            host: "*"
          # consumer group
          - resource:
              type: group
              name: connect-cluster
              patternType: literal
            operation: Read
            host: "*"
  2. Create or update the resource.

    kubectl apply -f KAFKA-USER-CONFIG-FILE

2.2.4. Performing a restart of a Kafka connector

This procedure describes how to manually trigger a restart of a Kafka connector by using a Kubernetes annotation.

Prerequisites
  • The Cluster Operator is running.

Procedure
  1. Find the name of the KafkaConnector custom resource that controls the Kafka connector you want to restart:

    kubectl get KafkaConnector
  2. To restart the connector, annotate the KafkaConnector resource in Kubernetes. For example, using kubectl annotate:

    kubectl annotate KafkaConnector KAFKACONNECTOR-NAME strimzi.io/restart=true
  3. Wait for the next reconciliation to occur (every two minutes by default).

    The Kafka connector is restarted, as long as the annotation was detected by the reconciliation process. When Kafka Connect accepts the restart request, the annotation is removed from the KafkaConnector custom resource.

Additional resources

2.2.5. Performing a restart of a Kafka connector task

This procedure describes how to manually trigger a restart of a Kafka connector task by using a Kubernetes annotation.

Prerequisites
  • The Cluster Operator is running.

Procedure
  1. Find the name of the KafkaConnector custom resource that controls the Kafka connector task you want to restart:

    kubectl get KafkaConnector
  2. Find the ID of the task to be restarted from the KafkaConnector custom resource. Task IDs are non-negative integers, starting from 0.

    kubectl describe KafkaConnector KAFKACONNECTOR-NAME
  3. To restart the connector task, annotate the KafkaConnector resource in Kubernetes. For example, using kubectl annotate to restart task 0:

    kubectl annotate KafkaConnector KAFKACONNECTOR-NAME strimzi.io/restart-task=0
  4. Wait for the next reconciliation to occur (every two minutes by default).

    The Kafka connector task is restarted, as long as the annotation was detected by the reconciliation process. When Kafka Connect accepts the restart request, the annotation is removed from the KafkaConnector custom resource.

Additional resources

2.2.6. Migrating from Kafka Connect with S2I to Kafka Connect

Support for Kafka Connect with S2I and the KafkaConnectS2I resource is deprecated. This follows the introduction of build configuration properties to the KafkaConnect resource, which are used to build a container image with the connector plugins you require for your data connections automatically.

This procedure describes how to migrate your Kafka Connect with S2I instance to a standard Kafka Connect instance. To do this, you configure a new KafkaConnect custom resource to replace the KafkaConnectS2I resource, which is then deleted.

Warning
The migration process involves downtime from the moment the KafkaConnectS2I instance is deleted until the new KafkaConnect instance has been successfully deployed. During this time, connectors will not be running and processing data. However, after the changeover they should continue from the point at which they stopped.
Prerequisites
  • Kafka Connect with S2I is deployed using a KafkaConnectS2I configuration

  • Kafka Connect with S2I is using an image with connectors added using an S2I build

  • Sink and source connector instances were created using KafkaConnector resources or the Kafka Connect REST API

Procedure
  1. Create a new KafkaConnect custom resource using the same name as the name used for the KafkaconnectS2I resource.

  2. Copy the KafkaConnectS2I resource properties to the KafkaConnect resource.

  3. If specified, make sure you use the same spec.config properties:

    • group.id

    • offset.storage.topic

    • config.storage.topic

    • status.storage.topic

      If these properties are not specified, defaults are used. In which case, leave them out of the KafkaConnect resource configuration as well.

    Now add configuration specific to the KafkaConnect resource to the new resource.

  4. Add build configuration to configure all the connectors and other libraries you want to add to the Kafka Connect deployment.

    Note
    Alternatively, you can build a new image with connectors manually, and specify it using the .spec.image property.
  5. Delete the old KafkaConnectS2I resource:

    kubectl delete -f MY-KAFKA-CONNECT-S2I-CONFIG-FILE

    Replace MY-KAFKA-CONNECT-S2I-CONFIG-FILE with the name of the file containing your KafkaConnectS2I resource configuration.

    Alternatively, you can specify the name of the resource:

    kubectl delete kafkaconnects2i MY-KAFKA-CONNECT-S2I

    Replace MY-KAFKA-CONNECT-S2I with the name of the KafkaConnectS2I resource.

    Wait until the Kafka Connect with S2I deployment and pods are deleted.

    Warning
    No other resources must be deleted.
  6. Deploy the new KafkaConnect resource:

    kubectl apply -f MY-KAFKA-CONNECT-CONFIG-FILE

    Replace MY-KAFKA-CONNECT-CONFIG-FILE with the name of the file containing your new KafkaConnect resource configuration.

    Wait until the new image is built, the deployment is created, and the pods have started.

  7. If you are using KafkaConnector resources for managing Kafka Connect connectors, check that all expected connectors are present and are running:

    kubectl get kctr --selector strimzi.io/cluster=MY-KAFKA-CONNECT-CLUSTER -o name

    Replace MY-KAFKA-CONNECT-CLUSTER with the name of your Kafka Connect cluster.

    Connectors automatically recover through Kafka Connect storage. Even if you are using the Kafka Connect REST API to manage them, you should not need to recreate them manually.

2.2.7. List of Kafka Connect cluster resources

The following resources are created by the Cluster Operator in the Kubernetes cluster:

connect-cluster-name-connect

Deployment which is in charge to create the Kafka Connect worker node pods.

connect-cluster-name-connect-api

Service which exposes the REST interface for managing the Kafka Connect cluster.

connect-cluster-name-config

ConfigMap which contains the Kafka Connect ancillary configuration and is mounted as a volume by the Kafka broker pods.

connect-cluster-name-connect

Pod Disruption Budget configured for the Kafka Connect worker nodes.

2.2.8. List of Kafka Connect (S2I) cluster resources

The following resources are created by the Cluster Operator in the Kubernetes cluster:

connect-cluster-name-connect-source

ImageStream which is used as the base image for the newly-built Docker images.

connect-cluster-name-connect

BuildConfig which is responsible for building the new Kafka Connect Docker images.

connect-cluster-name-connect

ImageStream where the newly built Docker images will be pushed.

connect-cluster-name-connect

DeploymentConfig which is in charge of creating the Kafka Connect worker node pods.

connect-cluster-name-connect-api

Service which exposes the REST interface for managing the Kafka Connect cluster.

connect-cluster-name-config

ConfigMap which contains the Kafka Connect ancillary configuration and is mounted as a volume by the Kafka broker pods.

connect-cluster-name-connect

Pod Disruption Budget configured for the Kafka Connect worker nodes.

2.3. Kafka MirrorMaker cluster configuration

This chapter describes how to configure a Kafka MirrorMaker deployment in your Strimzi cluster to replicate data between Kafka clusters.

You can use Strimzi with MirrorMaker or MirrorMaker 2.0. MirrorMaker 2.0 is the latest version, and offers a more efficient way to mirror data between Kafka clusters.

If you are using MirrorMaker, you configure the KafkaMirrorMaker resource.

The following procedure shows how the resource is configured:

The full schema of the KafkaMirrorMaker resource is described in the KafkaMirrorMaker schema reference.

2.3.1. Configuring Kafka MirrorMaker

Use the properties of the KafkaMirrorMaker resource to configure your Kafka MirrorMaker deployment.

You can configure access control for producers and consumers using TLS or SASL authentication. This procedure shows a configuration that uses TLS encryption and authentication on the consumer and producer side.

Prerequisites
  • See the Deploying and Upgrading Strimzi guide for instructions on running a:

  • Source and target Kafka clusters must be available

Procedure
  1. Edit the spec properties for the KafkaMirrorMaker resource.

    The properties you can configure are shown in this example configuration:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaMirrorMaker
    metadata:
      name: my-mirror-maker
    spec:
      replicas: 3 (1)
      consumer:
        bootstrapServers: my-source-cluster-kafka-bootstrap:9092 (2)
        groupId: "my-group" (3)
        numStreams: 2 (4)
        offsetCommitInterval: 120000 (5)
        tls: (6)
          trustedCertificates:
          - secretName: my-source-cluster-ca-cert
            certificate: ca.crt
        authentication: (7)
          type: tls
          certificateAndKey:
            secretName: my-source-secret
            certificate: public.crt
            key: private.key
        config: (8)
          max.poll.records: 100
          receive.buffer.bytes: 32768
          ssl.cipher.suites: "TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384" (9)
          ssl.enabled.protocols: "TLSv1.2"
          ssl.protocol: "TLSv1.2"
          ssl.endpoint.identification.algorithm: HTTPS (10)
      producer:
        bootstrapServers: my-target-cluster-kafka-bootstrap:9092
        abortOnSendFailure: false (11)
        tls:
          trustedCertificates:
          - secretName: my-target-cluster-ca-cert
            certificate: ca.crt
        authentication:
          type: tls
          certificateAndKey:
            secretName: my-target-secret
            certificate: public.crt
            key: private.key
        config:
          compression.type: gzip
          batch.size: 8192
          ssl.cipher.suites: "TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384" (12)
          ssl.enabled.protocols: "TLSv1.2"
          ssl.protocol: "TLSv1.2"
          ssl.endpoint.identification.algorithm: HTTPS (13)
      include: "my-topic|other-topic" (14)
      resources: (15)
        requests:
          cpu: "1"
          memory: 2Gi
        limits:
          cpu: "2"
          memory: 2Gi
      logging: (16)
        type: inline
        loggers:
          mirrormaker.root.logger: "INFO"
      readinessProbe: (17)
        initialDelaySeconds: 15
        timeoutSeconds: 5
      livenessProbe:
        initialDelaySeconds: 15
        timeoutSeconds: 5
      metricsConfig: (18)
       type: jmxPrometheusExporter
       valueFrom:
         configMapKeyRef:
           name: my-config-map
           key: my-key
      jvmOptions: (19)
        "-Xmx": "1g"
        "-Xms": "1g"
      image: my-org/my-image:latest (20)
      template: (21)
        pod:
          affinity:
            podAntiAffinity:
              requiredDuringSchedulingIgnoredDuringExecution:
                - labelSelector:
                    matchExpressions:
                      - key: application
                        operator: In
                        values:
                          - postgresql
                          - mongodb
                  topologyKey: "kubernetes.io/hostname"
        connectContainer: (22)
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
      tracing: (23)
        type: jaeger
    1. The number of replica nodes.

    2. Bootstrap servers for consumer and producer.

    3. Group ID for the consumer.

    4. The number of consumer streams.

    5. The offset auto-commit interval in milliseconds.

    6. TLS encryption with key names under which TLS certificates are stored in X.509 format for consumer or producer. If certificates are stored in the same secret, it can be listed multiple times.

    7. Authentication for consumer or producer, using the TLS mechanism, as shown here, using OAuth bearer tokens, or a SASL-based SCRAM-SHA-512 or PLAIN mechanism.

    8. Kafka configuration options for consumer and producer.

    9. SSL properties for external listeners to run with a specific cipher suite for a TLS version.

    10. Hostname verification is enabled by setting to HTTPS. An empty string disables the verification.

    11. If the abortOnSendFailure property is set to true, Kafka MirrorMaker will exit and the container will restart following a send failure for a message.

    12. SSL properties for external listeners to run with a specific cipher suite for a TLS version.

    13. Hostname verification is enabled by setting to HTTPS. An empty string disables the verification.

    14. A included topics mirrored from source to target Kafka cluster.

    15. Requests for reservation of supported resources, currently cpu and memory, and limits to specify the maximum resources that can be consumed.

    16. Specified loggers and log levels added directly (inline) or indirectly (external) through a ConfigMap. A custom ConfigMap must be placed under the log4j.properties or log4j2.properties key. MirrorMaker has a single logger called mirrormaker.root.logger. You can set the log level to INFO, ERROR, WARN, TRACE, DEBUG, FATAL or OFF.

    17. Healthchecks to know when to restart a container (liveness) and when a container can accept traffic (readiness).

    18. Prometheus metrics, which are enabled by referencing a ConfigMap containing configuration for the Prometheus JMX exporter in this example. You can enable metrics without further configuration using a reference to a ConfigMap containing an empty file under metricsConfig.valueFrom.configMapKeyRef.key.

    19. JVM configuration options to optimize performance for the Virtual Machine (VM) running Kafka MirrorMaker.

    20. ADVANCED OPTION: Container image configuration, which is recommended only in special situations.

    21. Template customization. Here a pod is scheduled with anti-affinity, so the pod is not scheduled on nodes with the same hostname.

    22. Environment variables are also set for distributed tracing using Jaeger.

    23. Distributed tracing is enabled for Jaeger.

    Warning
    With the abortOnSendFailure property set to false, the producer attempts to send the next message in a topic. The original message might be lost, as there is no attempt to resend a failed message.
  2. Create or update the resource:

    kubectl apply -f <your-file>

2.3.2. List of Kafka MirrorMaker cluster resources

The following resources are created by the Cluster Operator in the Kubernetes cluster:

<mirror-maker-name>-mirror-maker

Deployment which is responsible for creating the Kafka MirrorMaker pods.

<mirror-maker-name>-config

ConfigMap which contains ancillary configuration for the Kafka MirrorMaker, and is mounted as a volume by the Kafka broker pods.

<mirror-maker-name>-mirror-maker

Pod Disruption Budget configured for the Kafka MirrorMaker worker nodes.

2.4. Kafka MirrorMaker 2.0 cluster configuration

This section describes how to configure a Kafka MirrorMaker 2.0 deployment in your Strimzi cluster.

MirrorMaker 2.0 is used to replicate data between two or more active Kafka clusters, within or across data centers.

Data replication across clusters supports scenarios that require:

  • Recovery of data in the event of a system failure

  • Aggregation of data for analysis

  • Restriction of data access to a specific cluster

  • Provision of data at a specific location to improve latency

If you are using MirrorMaker 2.0, you configure the KafkaMirrorMaker2 resource.

MirrorMaker 2.0 introduces an entirely new way of replicating data between clusters.

As a result, the resource configuration differs from the previous version of MirrorMaker. If you choose to use MirrorMaker 2.0, there is currently no legacy support, so any resources must be manually converted into the new format.

How MirrorMaker 2.0 replicates data is described here:

The following procedure shows how the resource is configured for MirrorMaker 2.0:

The full schema of the KafkaMirrorMaker2 resource is described in the KafkaMirrorMaker2 schema reference.

2.4.1. MirrorMaker 2.0 data replication

MirrorMaker 2.0 consumes messages from a source Kafka cluster and writes them to a target Kafka cluster.

MirrorMaker 2.0 uses:

  • Source cluster configuration to consume data from the source cluster

  • Target cluster configuration to output data to the target cluster

MirrorMaker 2.0 is based on the Kafka Connect framework, connectors managing the transfer of data between clusters. A MirrorMaker 2.0 MirrorSourceConnector replicates topics from a source cluster to a target cluster.

The process of mirroring data from one cluster to another cluster is asynchronous. The recommended pattern is for messages to be produced locally alongside the source Kafka cluster, then consumed remotely close to the target Kafka cluster.

MirrorMaker 2.0 can be used with more than one source cluster.

MirrorMaker 2.0 replication
Figure 1. Replication across two clusters

By default, a check for new topics in the source cluster is made every 10 minutes. You can change the frequency by adding refresh.topics.interval.seconds to the source connector configuration. However, increasing the frequency of the operation might affect overall performance.

2.4.2. Cluster configuration

You can use MirrorMaker 2.0 in active/passive or active/active cluster configurations.

  • In an active/active configuration, both clusters are active and provide the same data simultaneously, which is useful if you want to make the same data available locally in different geographical locations.

  • In an active/passive configuration, the data from an active cluster is replicated in a passive cluster, which remains on standby, for example, for data recovery in the event of system failure.

The expectation is that producers and consumers connect to active clusters only.

A MirrorMaker 2.0 cluster is required at each target destination.

Bidirectional replication (active/active)

The MirrorMaker 2.0 architecture supports bidirectional replication in an active/active cluster configuration.

Each cluster replicates the data of the other cluster using the concept of source and remote topics. As the same topics are stored in each cluster, remote topics are automatically renamed by MirrorMaker 2.0 to represent the source cluster. The name of the originating cluster is prepended to the name of the topic.

MirrorMaker 2.0 bidirectional architecture
Figure 2. Topic renaming

By flagging the originating cluster, topics are not replicated back to that cluster.

The concept of replication through remote topics is useful when configuring an architecture that requires data aggregation. Consumers can subscribe to source and remote topics within the same cluster, without the need for a separate aggregation cluster.

Unidirectional replication (active/passive)

The MirrorMaker 2.0 architecture supports unidirectional replication in an active/passive cluster configuration.

You can use an active/passive cluster configuration to make backups or migrate data to another cluster. In this situation, you might not want automatic renaming of remote topics.

You can override automatic renaming by adding IdentityReplicationPolicy to the source connector configuration. With this configuration applied, topics retain their original names.

Topic configuration synchronization

Topic configuration is automatically synchronized between source and target clusters. By synchronizing configuration properties, the need for rebalancing is reduced.

Data integrity

MirrorMaker 2.0 monitors source topics and propagates any configuration changes to remote topics, checking for and creating missing partitions. Only MirrorMaker 2.0 can write to remote topics.

Offset tracking

MirrorMaker 2.0 tracks offsets for consumer groups using internal topics.

  • The offset sync topic maps the source and target offsets for replicated topic partitions from record metadata

  • The checkpoint topic maps the last committed offset in the source and target cluster for replicated topic partitions in each consumer group

Offsets for the checkpoint topic are tracked at predetermined intervals through configuration. Both topics enable replication to be fully restored from the correct offset position on failover.

MirrorMaker 2.0 uses its MirrorCheckpointConnector to emit checkpoints for offset tracking.

Synchronizing consumer group offsets

The __consumer_offsets topic stores information on committed offsets, for each consumer group. Offset synchronization periodically transfers the consumer offsets for the consumer groups of a source cluster into the consumer offsets topic of a target cluster.

Offset synchronization is particularly useful in an active/passive configuration. If the active cluster goes down, consumer applications can switch to the passive (standby) cluster and pick up from the last transferred offset position.

To use topic offset synchronization:

  • Enable the synchronization by adding sync.group.offsets.enabled to the checkpoint connector configuration, and setting the property to true. Synchronization is disabled by default.

  • Add the IdentityReplicationPolicy to the source and checkpoint connector configuration so that topics in the target cluster retain their original names.

For topic offset synchronization to work, consumer groups in the target cluster cannot use the same ids as groups in the source cluster.

If enabled, the synchronization of offsets from the source cluster is made periodically. You can change the frequency by adding sync.group.offsets.interval.seconds and emit.checkpoints.interval.seconds to the checkpoint connector configuration. The properties specify the frequency in seconds that the consumer group offsets are synchronized, and the frequency of checkpoints emitted for offset tracking. The default for both properties is 60 seconds. You can also change the frequency of checks for new consumer groups using the refresh.groups.interval.seconds property, which is performed every 10 minutes by default.

Because the synchronization is time-based, any switchover by consumers to a passive cluster will likely result in some duplication of messages.

Connectivity checks

A heartbeat internal topic checks connectivity between clusters.

The heartbeat topic is replicated from the source cluster.

Target clusters use the topic to check:

  • The connector managing connectivity between clusters is running

  • The source cluster is available

MirrorMaker 2.0 uses its MirrorHeartbeatConnector to emit heartbeats that perform these checks.

2.4.3. ACL rules synchronization

ACL access to remote topics is possible if you are not using the User Operator.

If AclAuthorizer is being used, without the User Operator, ACL rules that manage access to brokers also apply to remote topics. Users that can read a source topic can read its remote equivalent.

Note
OAuth 2.0 authorization does not support access to remote topics in this way.

2.4.4. Synchronizing data between Kafka clusters using MirrorMaker 2.0

Use MirrorMaker 2.0 to synchronize data between Kafka clusters through configuration.

The configuration must specify:

  • Each Kafka cluster

  • Connection information for each cluster, including TLS authentication

  • The replication flow and direction

    • Cluster to cluster

    • Topic to topic

Use the properties of the KafkaMirrorMaker2 resource to configure your Kafka MirrorMaker 2.0 deployment.

Note
The previous version of MirrorMaker continues to be supported. If you wish to use the resources configured for the previous version, they must be updated to the format supported by MirrorMaker 2.0.

MirrorMaker 2.0 provides default configuration values for properties such as replication factors. A minimal configuration, with defaults left unchanged, would be something like this example:

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaMirrorMaker2
metadata:
  name: my-mirror-maker2
spec:
  version: 2.8.0
  connectCluster: "my-cluster-target"
  clusters:
  - alias: "my-cluster-source"
    bootstrapServers: my-cluster-source-kafka-bootstrap:9092
  - alias: "my-cluster-target"
    bootstrapServers: my-cluster-target-kafka-bootstrap:9092
  mirrors:
  - sourceCluster: "my-cluster-source"
    targetCluster: "my-cluster-target"
    sourceConnector: {}

You can configure access control for source and target clusters using TLS or SASL authentication. This procedure shows a configuration that uses TLS encryption and authentication for the source and target cluster.

Prerequisites
  • See the Deploying and Upgrading Strimzi guide for instructions on running a:

  • Source and target Kafka clusters must be available

Procedure
  1. Edit the spec properties for the KafkaMirrorMaker2 resource.

    The properties you can configure are shown in this example configuration:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaMirrorMaker2
    metadata:
      name: my-mirror-maker2
    spec:
      version: 2.8.0 (1)
      replicas: 3 (2)
      connectCluster: "my-cluster-target" (3)
      clusters: (4)
      - alias: "my-cluster-source" (5)
        authentication: (6)
          certificateAndKey:
            certificate: source.crt
            key: source.key
            secretName: my-user-source
          type: tls
        bootstrapServers: my-cluster-source-kafka-bootstrap:9092 (7)
        tls: (8)
          trustedCertificates:
          - certificate: ca.crt
            secretName: my-cluster-source-cluster-ca-cert
      - alias: "my-cluster-target" (9)
        authentication: (10)
          certificateAndKey:
            certificate: target.crt
            key: target.key
            secretName: my-user-target
          type: tls
        bootstrapServers: my-cluster-target-kafka-bootstrap:9092 (11)
        config: (12)
          config.storage.replication.factor: 1
          offset.storage.replication.factor: 1
          status.storage.replication.factor: 1
          ssl.cipher.suites: "TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384" (13)
          ssl.enabled.protocols: "TLSv1.2"
          ssl.protocol: "TLSv1.2"
          ssl.endpoint.identification.algorithm: HTTPS (14)
        tls: (15)
          trustedCertificates:
          - certificate: ca.crt
            secretName: my-cluster-target-cluster-ca-cert
      mirrors: (16)
      - sourceCluster: "my-cluster-source" (17)
        targetCluster: "my-cluster-target" (18)
        sourceConnector: (19)
          tasksMax: 10 (20)
          config:
            replication.factor: 1 (21)
            offset-syncs.topic.replication.factor: 1 (22)
            sync.topic.acls.enabled: "false" (23)
            refresh.topics.interval.seconds: 60 (24)
            replication.policy.separator: "" (25)
            replication.policy.class: "io.strimzi.kafka.connect.mirror.IdentityReplicationPolicy" (26)
        heartbeatConnector: (27)
          config:
            heartbeats.topic.replication.factor: 1 (28)
        checkpointConnector: (29)
          config:
            checkpoints.topic.replication.factor: 1 (30)
            refresh.groups.interval.seconds: 600 (31)
            sync.group.offsets.enabled: true (32)
            sync.group.offsets.interval.seconds: 60 (33)
            emit.checkpoints.interval.seconds: 60 (34)
            replication.policy.class: "io.strimzi.kafka.connect.mirror.IdentityReplicationPolicy"
        topicsPattern: ".*" (35)
        groupsPattern: "group1|group2|group3" (36)
      resources: (37)
        requests:
          cpu: "1"
          memory: 2Gi
        limits:
          cpu: "2"
          memory: 2Gi
      logging: (38)
        type: inline
        loggers:
          connect.root.logger.level: "INFO"
      readinessProbe: (39)
        initialDelaySeconds: 15
        timeoutSeconds: 5
      livenessProbe:
        initialDelaySeconds: 15
        timeoutSeconds: 5
      jvmOptions: (40)
        "-Xmx": "1g"
        "-Xms": "1g"
      image: my-org/my-image:latest (41)
      template: (42)
        pod:
          affinity:
            podAntiAffinity:
              requiredDuringSchedulingIgnoredDuringExecution:
                - labelSelector:
                    matchExpressions:
                      - key: application
                        operator: In
                        values:
                          - postgresql
                          - mongodb
                  topologyKey: "kubernetes.io/hostname"
        connectContainer: (43)
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
      tracing:
        type: jaeger (44)
      externalConfiguration: (45)
        env:
          - name: AWS_ACCESS_KEY_ID
            valueFrom:
              secretKeyRef:
                name: aws-creds
                key: awsAccessKey
          - name: AWS_SECRET_ACCESS_KEY
            valueFrom:
              secretKeyRef:
                name: aws-creds
                key: awsSecretAccessKey
    1. The Kafka Connect and Mirror Maker 2.0 version, which will always be the same.

    2. The number of replica nodes.

    3. Kafka cluster alias for Kafka Connect, which must specify the target Kafka cluster. The Kafka cluster is used by Kafka Connect for its internal topics.

    4. Specification for the Kafka clusters being synchronized.

    5. Cluster alias for the source Kafka cluster.

    6. Authentication for the source cluster, using the TLS mechanism, as shown here, using OAuth bearer tokens, or a SASL-based SCRAM-SHA-512 or PLAIN mechanism.

    7. Bootstrap server for connection to the source Kafka cluster.

    8. TLS encryption with key names under which TLS certificates are stored in X.509 format for the source Kafka cluster. If certificates are stored in the same secret, it can be listed multiple times.

    9. Cluster alias for the target Kafka cluster.

    10. Authentication for the target Kafka cluster is configured in the same way as for the source Kafka cluster.

    11. Bootstrap server for connection to the target Kafka cluster.

    12. Kafka Connect configuration. Standard Apache Kafka configuration may be provided, restricted to those properties not managed directly by Strimzi.

    13. SSL properties for external listeners to run with a specific cipher suite for a TLS version.

    14. Hostname verification is enabled by setting to HTTPS. An empty string disables the verification.

    15. TLS encryption for the target Kafka cluster is configured in the same way as for the source Kafka cluster.

    16. MirrorMaker 2.0 connectors.

    17. Cluster alias for the source cluster used by the MirrorMaker 2.0 connectors.

    18. Cluster alias for the target cluster used by the MirrorMaker 2.0 connectors.

    19. Configuration for the MirrorSourceConnector that creates remote topics. The config overrides the default configuration options.

    20. The maximum number of tasks that the connector may create. Tasks handle the data replication and run in parallel. If the infrastructure supports the processing overhead, increasing this value can improve throughput. Kafka Connect distributes the tasks between members of the cluster. If there are more tasks than workers, workers are assigned multiple tasks. For sink connectors, aim to have one task for each topic partition consumed. For source connectors, the number of tasks that can run in parallel may also depend on the external system. The connector creates fewer than the maximum number of tasks if it cannot achieve the parallelism.

    21. Replication factor for mirrored topics created at the target cluster.

    22. Replication factor for the MirrorSourceConnector offset-syncs internal topic that maps the offsets of the source and target clusters.

    23. When ACL rules synchronization is enabled, ACLs are applied to synchronized topics. The default is true.

    24. Optional setting to change the frequency of checks for new topics. The default is for a check every 10 minutes.

    25. Defines the separator used for the renaming of remote topics.

    26. Adds a policy that overrides the automatic renaming of remote topics. Instead of prepending the name with the name of the source cluster, the topic retains its original name. This optional setting is useful for active/passive backups and data migration. To configure topic offset synchronization, this property must also be set for the checkpointConnector.config.

    27. Configuration for the MirrorHeartbeatConnector that performs connectivity checks. The config overrides the default configuration options.

    28. Replication factor for the heartbeat topic created at the target cluster.

    29. Configuration for the MirrorCheckpointConnector that tracks offsets. The config overrides the default configuration options.

    30. Replication factor for the checkpoints topic created at the target cluster.

    31. Optional setting to change the frequency of checks for new consumer groups. The default is for a check every 10 minutes.

    32. Optional setting to synchronize consumer group offsets, which is useful for recovery in an active/passive configuration. Synchronization is not enabled by default.

    33. If the synchronization of consumer group offsets is enabled, you can adjust the frequency of the synchronization.

    34. Adjusts the frequency of checks for offset tracking. If you change the frequency of offset synchronization, you might also need to adjust the frequency of these checks.

    35. Topic replication from the source cluster defined as regular expression patterns. Here we request all topics.

    36. Consumer group replication from the source cluster defined as regular expression patterns. Here we request three consumer groups by name. You can use comma-separated lists.

    37. Requests for reservation of supported resources, currently cpu and memory, and limits to specify the maximum resources that can be consumed.

    38. Specified Kafka Connect loggers and log levels added directly (inline) or indirectly (external) through a ConfigMap. A custom ConfigMap must be placed under the log4j.properties or log4j2.properties key. For the Kafka Connect log4j.rootLogger logger, you can set the log level to INFO, ERROR, WARN, TRACE, DEBUG, FATAL or OFF.

    39. Healthchecks to know when to restart a container (liveness) and when a container can accept traffic (readiness).

    40. JVM configuration options to optimize performance for the Virtual Machine (VM) running Kafka MirrorMaker.

    41. ADVANCED OPTION: Container image configuration, which is recommended only in special situations.

    42. Template customization. Here a pod is scheduled with anti-affinity, so the pod is not scheduled on nodes with the same hostname.

    43. Environment variables are also set for distributed tracing using Jaeger.

    44. Distributed tracing is enabled for Jaeger.

    45. External configuration for a Kubernetes Secret mounted to Kafka MirrorMaker as an environment variable. You can also use the Kubernetes Configuration Provider to load configuration values from external sources.

  2. Create or update the resource:

    kubectl apply -f MIRRORMAKER-CONFIGURATION-FILE

2.4.5. Performing a restart of a Kafka MirrorMaker 2.0 connector

This procedure describes how to manually trigger a restart of a Kafka MirrorMaker 2.0 connector by using a Kubernetes annotation.

Prerequisites
  • The Cluster Operator is running.

Procedure
  1. Find the name of the KafkaMirrorMaker2 custom resource that controls the Kafka MirrorMaker 2.0 connector you want to restart:

    kubectl get KafkaMirrorMaker2
  2. Find the name of the Kafka MirrorMaker 2.0 connector to be restarted from the KafkaMirrorMaker2 custom resource.

    kubectl describe KafkaMirrorMaker2 KAFKAMIRRORMAKER-2-NAME
  3. To restart the connector, annotate the KafkaMirrorMaker2 resource in Kubernetes. In this example, kubectl annotate restarts a connector named my-source->my-target.MirrorSourceConnector:

    kubectl annotate KafkaMirrorMaker2 KAFKAMIRRORMAKER-2-NAME "strimzi.io/restart-connector=my-source->my-target.MirrorSourceConnector"
  4. Wait for the next reconciliation to occur (every two minutes by default).

    The Kafka MirrorMaker 2.0 connector is restarted, as long as the annotation was detected by the reconciliation process. When the restart request is accepted, the annotation is removed from the KafkaMirrorMaker2 custom resource.

2.4.6. Performing a restart of a Kafka MirrorMaker 2.0 connector task

This procedure describes how to manually trigger a restart of a Kafka MirrorMaker 2.0 connector task by using a Kubernetes annotation.

Prerequisites
  • The Cluster Operator is running.

Procedure
  1. Find the name of the KafkaMirrorMaker2 custom resource that controls the Kafka MirrorMaker 2.0 connector you want to restart:

    kubectl get KafkaMirrorMaker2
  2. Find the name of the Kafka MirrorMaker 2.0 connector and the ID of the task to be restarted from the KafkaMirrorMaker2 custom resource. Task IDs are non-negative integers, starting from 0.

    kubectl describe KafkaMirrorMaker2 KAFKAMIRRORMAKER-2-NAME
  3. To restart the connector task, annotate the KafkaMirrorMaker2 resource in Kubernetes. In this example, kubectl annotate restarts task 0 of a connector named my-source->my-target.MirrorSourceConnector:

    kubectl annotate KafkaMirrorMaker2 KAFKAMIRRORMAKER-2-NAME "strimzi.io/restart-connector-task=my-source->my-target.MirrorSourceConnector:0"
  4. Wait for the next reconciliation to occur (every two minutes by default).

    The Kafka MirrorMaker 2.0 connector task is restarted, as long as the annotation was detected by the reconciliation process. When the restart task request is accepted, the annotation is removed from the KafkaMirrorMaker2 custom resource.

2.5. Kafka Bridge cluster configuration

This section describes how to configure a Kafka Bridge deployment in your Strimzi cluster.

Kafka Bridge provides an API for integrating HTTP-based clients with a Kafka cluster.

If you are using the Kafka Bridge, you configure the KafkaBridge resource.

The full schema of the KafkaBridge resource is described in KafkaBridge schema reference.

2.5.1. Configuring the Kafka Bridge

Use the Kafka Bridge to make HTTP-based requests to the Kafka cluster.

Use the properties of the KafkaBridge resource to configure your Kafka Bridge deployment.

In order to prevent issues arising when client consumer requests are processed by different Kafka Bridge instances, address-based routing must be employed to ensure that requests are routed to the right Kafka Bridge instance. Additionally, each independent Kafka Bridge instance must have a replica. A Kafka Bridge instance has its own state which is not shared with another instances.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

See the Deploying and Upgrading Strimzi guide for instructions on running a:

Procedure
  1. Edit the spec properties for the KafkaBridge resource.

    The properties you can configure are shown in this example configuration:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaBridge
    metadata:
      name: my-bridge
    spec:
      replicas: 3 (1)
      bootstrapServers: my-cluster-kafka-bootstrap:9092 (2)
      tls: (3)
        trustedCertificates:
          - secretName: my-cluster-cluster-cert
            certificate: ca.crt
          - secretName: my-cluster-cluster-cert
            certificate: ca2.crt
      authentication: (4)
        type: tls
        certificateAndKey:
          secretName: my-secret
          certificate: public.crt
          key: private.key
      http: (5)
        port: 8080
        cors: (6)
          allowedOrigins: "https://strimzi.io"
          allowedMethods: "GET,POST,PUT,DELETE,OPTIONS,PATCH"
      consumer: (7)
        config:
          auto.offset.reset: earliest
      producer: (8)
        config:
          delivery.timeout.ms: 300000
      resources: (9)
        requests:
          cpu: "1"
          memory: 2Gi
        limits:
          cpu: "2"
          memory: 2Gi
      logging: (10)
        type: inline
        loggers:
          logger.bridge.level: "INFO"
          # enabling DEBUG just for send operation
          logger.send.name: "http.openapi.operation.send"
          logger.send.level: "DEBUG"
      jvmOptions: (11)
        "-Xmx": "1g"
        "-Xms": "1g"
      readinessProbe: (12)
        initialDelaySeconds: 15
        timeoutSeconds: 5
      livenessProbe:
        initialDelaySeconds: 15
        timeoutSeconds: 5
      image: my-org/my-image:latest (13)
      template: (14)
        pod:
          affinity:
            podAntiAffinity:
              requiredDuringSchedulingIgnoredDuringExecution:
                - labelSelector:
                    matchExpressions:
                      - key: application
                        operator: In
                        values:
                          - postgresql
                          - mongodb
                  topologyKey: "kubernetes.io/hostname"
        bridgeContainer: (15)
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
    1. The number of replica nodes.

    2. Bootstrap server for connection to the target Kafka cluster.

    3. TLS encryption with key names under which TLS certificates are stored in X.509 format for the source Kafka cluster. If certificates are stored in the same secret, it can be listed multiple times.

    4. Authentication for the Kafka Bridge cluster, using the TLS mechanism, as shown here, using OAuth bearer tokens, or a SASL-based SCRAM-SHA-512 or PLAIN mechanism. By default, the Kafka Bridge connects to Kafka brokers without authentication.

    5. HTTP access to Kafka brokers.

    6. CORS access specifying selected resources and access methods. Additional HTTP headers in requests describe the origins that are permitted access to the Kafka cluster.

    7. Consumer configuration options.

    8. Producer configuration options.

    9. Requests for reservation of supported resources, currently cpu and memory, and limits to specify the maximum resources that can be consumed.

    10. Specified Kafka Bridge loggers and log levels added directly (inline) or indirectly (external) through a ConfigMap. A custom ConfigMap must be placed under the log4j.properties or log4j2.properties key. For the Kafka Bridge loggers, you can set the log level to INFO, ERROR, WARN, TRACE, DEBUG, FATAL or OFF.

    11. JVM configuration options to optimize performance for the Virtual Machine (VM) running the Kafka Bridge.

    12. Healthchecks to know when to restart a container (liveness) and when a container can accept traffic (readiness).

    13. ADVANCED OPTION: Container image configuration, which is recommended only in special situations.

    14. Template customization. Here a pod is scheduled with anti-affinity, so the pod is not scheduled on nodes with the same hostname.

    15. Environment variables are also set for distributed tracing using Jaeger.

  2. Create or update the resource:

    kubectl apply -f KAFKA-BRIDGE-CONFIG-FILE

2.5.2. List of Kafka Bridge cluster resources

The following resources are created by the Cluster Operator in the Kubernetes cluster:

bridge-cluster-name-bridge

Deployment which is in charge to create the Kafka Bridge worker node pods.

bridge-cluster-name-bridge-service

Service which exposes the REST interface of the Kafka Bridge cluster.

bridge-cluster-name-bridge-config

ConfigMap which contains the Kafka Bridge ancillary configuration and is mounted as a volume by the Kafka broker pods.

bridge-cluster-name-bridge

Pod Disruption Budget configured for the Kafka Bridge worker nodes.

2.6. Customizing Kubernetes resources

Strimzi creates several Kubernetes resources, such as Deployments, StatefulSets, Pods, and Services, which are managed by Strimzi operators. Only the operator that is responsible for managing a particular Kubernetes resource can change that resource. If you try to manually change an operator-managed Kubernetes resource, the operator will revert your changes back.

However, changing an operator-managed Kubernetes resource can be useful if you want to perform certain tasks, such as:

  • Adding custom labels or annotations that control how Pods are treated by Istio or other services

  • Managing how Loadbalancer-type Services are created by the cluster

You can make such changes using the template property in the Strimzi custom resources. The template property is supported in the following resources. The API reference provides more details about the customizable fields.

Kafka.spec.kafka

See KafkaClusterTemplate schema reference

Kafka.spec.zookeeper

See ZookeeperClusterTemplate schema reference

Kafka.spec.entityOperator

See EntityOperatorTemplate schema reference

Kafka.spec.kafkaExporter

See KafkaExporterTemplate schema reference

Kafka.spec.cruiseControl

See CruiseControlTemplate schema reference

Kafka.spec.jmxTrans

See JmxTransTemplate schema reference

KafkaConnect.spec

See KafkaConnectTemplate schema reference

KafkaConnectS2I.spec

See KafkaConnectTemplate schema reference

KafkaMirrorMaker.spec

See KafkaMirrorMakerTemplate schema reference

KafkaMirrorMaker2.spec

See KafkaConnectTemplate schema reference

KafkaBridge.spec

See KafkaBridgeTemplate schema reference

KafkaUser.spec

See KafkaUserTemplate schema reference

In the following example, the template property is used to modify the labels in a Kafka broker’s StatefulSet:

Example template customization
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
  labels:
    app: my-cluster
spec:
  kafka:
    # ...
    template:
      statefulset:
        metadata:
          labels:
            mylabel: myvalue
    # ...

2.6.1. Customizing the image pull policy

Strimzi allows you to customize the image pull policy for containers in all pods deployed by the Cluster Operator. The image pull policy is configured using the environment variable STRIMZI_IMAGE_PULL_POLICY in the Cluster Operator deployment. The STRIMZI_IMAGE_PULL_POLICY environment variable can be set to three different values:

Always

Container images are pulled from the registry every time the pod is started or restarted.

IfNotPresent

Container images are pulled from the registry only when they were not pulled before.

Never

Container images are never pulled from the registry.

The image pull policy can be currently customized only for all Kafka, Kafka Connect, and Kafka MirrorMaker clusters at once. Changing the policy will result in a rolling update of all your Kafka, Kafka Connect, and Kafka MirrorMaker clusters.

Additional resources

2.7. Configuring pod scheduling

When two applications are scheduled to the same Kubernetes node, both applications might use the same resources like disk I/O and impact performance. That can lead to performance degradation. Scheduling Kafka pods in a way that avoids sharing nodes with other critical workloads, using the right nodes or dedicated a set of nodes only for Kafka are the best ways how to avoid such problems.

2.7.1. Specifying affinity, tolerations, and topology spread constraints

Use affinity, tolerations and topology spread constraints to schedule the pods of kafka resources onto nodes. Affinity, tolerations and topology spread constraints are configured using the affinity, tolerations, and topologySpreadConstraint properties in following resources:

  • Kafka.spec.kafka.template.pod

  • Kafka.spec.zookeeper.template.pod

  • Kafka.spec.entityOperator.template.pod

  • KafkaConnect.spec.template.pod

  • KafkaConnectS2I.spec.template.pod

  • KafkaBridge.spec.template.pod

  • KafkaMirrorMaker.spec.template.pod

  • KafkaMirrorMaker2.spec.template.pod

The format of the affinity, tolerations, and topologySpreadConstraint properties follows the Kubernetes specification. The affinity configuration can include different types of affinity:

  • Pod affinity and anti-affinity

  • Node affinity

Note
On Kubernetes 1.16 and 1.17, the support for topologySpreadConstraint is disabled by default. In order to use topologySpreadConstraint, you have to enable the EvenPodsSpread feature gate in Kubernetes API server and scheduler.
Use pod anti-affinity to avoid critical applications sharing nodes

Use pod anti-affinity to ensure that critical applications are never scheduled on the same disk. When running a Kafka cluster, it is recommended to use pod anti-affinity to ensure that the Kafka brokers do not share nodes with other workloads, such as databases.

Use node affinity to schedule workloads onto specific nodes

The Kubernetes cluster usually consists of many different types of worker nodes. Some are optimized for CPU heavy workloads, some for memory, while other might be optimized for storage (fast local SSDs) or network. Using different nodes helps to optimize both costs and performance. To achieve the best possible performance, it is important to allow scheduling of Strimzi components to use the right nodes.

Kubernetes uses node affinity to schedule workloads onto specific nodes. Node affinity allows you to create a scheduling constraint for the node on which the pod will be scheduled. The constraint is specified as a label selector. You can specify the label using either the built-in node label like beta.kubernetes.io/instance-type or custom labels to select the right node.

Use node affinity and tolerations for dedicated nodes

Use taints to create dedicated nodes, then schedule Kafka pods on the dedicated nodes by configuring node affinity and tolerations.

Cluster administrators can mark selected Kubernetes nodes as tainted. Nodes with taints are excluded from regular scheduling and normal pods will not be scheduled to run on them. Only services which can tolerate the taint set on the node can be scheduled on it. The only other services running on such nodes will be system services such as log collectors or software defined networks.

Running Kafka and its components on dedicated nodes can have many advantages. There will be no other applications running on the same nodes which could cause disturbance or consume the resources needed for Kafka. That can lead to improved performance and stability.

2.7.2. Configuring pod anti-affinity to schedule each Kafka broker on a different worker node

Many Kafka brokers or ZooKeeper nodes can run on the same Kubernetes worker node. If the worker node fails, they will all become unavailable at the same time. To improve reliability, you can use podAntiAffinity configuration to schedule each Kafka broker or ZooKeeper node on a different Kubernetes worker node.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Edit the affinity property in the resource specifying the cluster deployment. To make sure that no worker nodes are shared by Kafka brokers or ZooKeeper nodes, use the strimzi.io/name label. Set the topologyKey to kubernetes.io/hostname to specify that the selected pods are not scheduled on nodes with the same hostname. This will still allow the same worker node to be shared by a single Kafka broker and a single ZooKeeper node. For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        template:
          pod:
            affinity:
              podAntiAffinity:
                requiredDuringSchedulingIgnoredDuringExecution:
                  - labelSelector:
                      matchExpressions:
                        - key: strimzi.io/name
                          operator: In
                          values:
                            - CLUSTER-NAME-kafka
                    topologyKey: "kubernetes.io/hostname"
        # ...
      zookeeper:
        # ...
        template:
          pod:
            affinity:
              podAntiAffinity:
                requiredDuringSchedulingIgnoredDuringExecution:
                  - labelSelector:
                      matchExpressions:
                        - key: strimzi.io/name
                          operator: In
                          values:
                            - CLUSTER-NAME-zookeeper
                    topologyKey: "kubernetes.io/hostname"
        # ...

    Where CLUSTER-NAME is the name of your Kafka custom resource.

  2. If you even want to make sure that a Kafka broker and ZooKeeper node do not share the same worker node, use the strimzi.io/cluster label. For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        template:
          pod:
            affinity:
              podAntiAffinity:
                requiredDuringSchedulingIgnoredDuringExecution:
                  - labelSelector:
                      matchExpressions:
                        - key: strimzi.io/cluster
                          operator: In
                          values:
                            - CLUSTER-NAME
                    topologyKey: "kubernetes.io/hostname"
        # ...
      zookeeper:
        # ...
        template:
          pod:
            affinity:
              podAntiAffinity:
                requiredDuringSchedulingIgnoredDuringExecution:
                  - labelSelector:
                      matchExpressions:
                        - key: strimzi.io/cluster
                          operator: In
                          values:
                            - CLUSTER-NAME
                    topologyKey: "kubernetes.io/hostname"
        # ...

    Where CLUSTER-NAME is the name of your Kafka custom resource.

  3. Create or update the resource.

    kubectl apply -f KAFKA-CONFIG-FILE

2.7.3. Configuring pod anti-affinity in Kafka components

Pod anti-affinity configuration helps with the stability and performance of Kafka brokers. By using podAntiAffinity, Kubernetes will not schedule Kafka brokers on the same nodes as other workloads. Typically, you want to avoid Kafka running on the same worker node as other network or storage intensive applications such as databases, storage or other messaging platforms.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Edit the affinity property in the resource specifying the cluster deployment. Use labels to specify the pods which should not be scheduled on the same nodes. The topologyKey should be set to kubernetes.io/hostname to specify that the selected pods should not be scheduled on nodes with the same hostname. For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        template:
          pod:
            affinity:
              podAntiAffinity:
                requiredDuringSchedulingIgnoredDuringExecution:
                  - labelSelector:
                      matchExpressions:
                        - key: application
                          operator: In
                          values:
                            - postgresql
                            - mongodb
                    topologyKey: "kubernetes.io/hostname"
        # ...
      zookeeper:
        # ...
  2. Create or update the resource.

    This can be done using kubectl apply:

    kubectl apply -f KAFKA-CONFIG-FILE

2.7.4. Configuring node affinity in Kafka components

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Label the nodes where Strimzi components should be scheduled.

    This can be done using kubectl label:

    kubectl label node NAME-OF-NODE node-type=fast-network

    Alternatively, some of the existing labels might be reused.

  2. Edit the affinity property in the resource specifying the cluster deployment. For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        template:
          pod:
            affinity:
              nodeAffinity:
                requiredDuringSchedulingIgnoredDuringExecution:
                  nodeSelectorTerms:
                    - matchExpressions:
                      - key: node-type
                        operator: In
                        values:
                        - fast-network
        # ...
      zookeeper:
        # ...
  3. Create or update the resource.

    This can be done using kubectl apply:

    kubectl apply -f KAFKA-CONFIG-FILE

2.7.5. Setting up dedicated nodes and scheduling pods on them

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Select the nodes which should be used as dedicated.

  2. Make sure there are no workloads scheduled on these nodes.

  3. Set the taints on the selected nodes:

    This can be done using kubectl taint:

    kubectl taint node NAME-OF-NODE dedicated=Kafka:NoSchedule
  4. Additionally, add a label to the selected nodes as well.

    This can be done using kubectl label:

    kubectl label node NAME-OF-NODE dedicated=Kafka
  5. Edit the affinity and tolerations properties in the resource specifying the cluster deployment.

    For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        template:
          pod:
            tolerations:
              - key: "dedicated"
                operator: "Equal"
                value: "Kafka"
                effect: "NoSchedule"
            affinity:
              nodeAffinity:
                requiredDuringSchedulingIgnoredDuringExecution:
                  nodeSelectorTerms:
                  - matchExpressions:
                    - key: dedicated
                      operator: In
                      values:
                      - Kafka
        # ...
      zookeeper:
        # ...
  6. Create or update the resource.

    This can be done using kubectl apply:

    kubectl apply -f KAFKA-CONFIG-FILE

2.8. Logging configuration

Configure logging levels in the custom resources of Kafka components and Strimzi Operators. You can specify the logging levels directly in the spec.logging property of the custom resource. Or you can define the logging properties in a ConfigMap that’s referenced in the custom resource using the configMapKeyRef property.

The advantages of using a ConfigMap are that the logging properties are maintained in one place and are accessible to more than one resource. You can also reuse the ConfigMap for more than one resource. If you are using a ConfigMap to specify loggers for Strimzi Operators, you can also append the logging specification to add filters.

You specify a logging type in your logging specification:

  • inline when specifying logging levels directly

  • external when referencing a ConfigMap

Example inline logging configuration
spec:
  # ...
  logging:
    type: inline
    loggers:
      kafka.root.logger.level: "INFO"
Example external logging configuration
spec:
  # ...
  logging:
    type: external
    valueFrom:
      configMapKeyRef:
        name: my-config-map
        key: my-config-map-key

Values for the name and key of the ConfigMap are mandatory. Default logging is used if the name or key is not set.

Additional resources

Kafka component logging

Operator logging

2.8.1. Creating a ConfigMap for logging

To use a ConfigMap to define logging properties, you create the ConfigMap and then reference it as part of the logging definition in the spec of a resource.

The ConfigMap must contain the appropriate logging configuration.

  • log4j.properties for Kafka components, ZooKeeper, and the Kafka Bridge

  • log4j2.properties for the Topic Operator and User Operator

The configuration must be placed under these properties.

In this procedure a ConfigMap defines a root logger for a Kafka resource.

Procedure
  1. Create the ConfigMap.

    You can create the ConfigMap as a YAML file or from a properties file.

    ConfigMap example with a root logger definition for Kafka:

    kind: ConfigMap
    apiVersion: v1
    metadata:
      name: logging-configmap
    data:
      log4j.properties:
        kafka.root.logger.level="INFO"

    If you are using a properties file, specify the file at the command line:

    kubectl create configmap logging-configmap --from-file=log4j.properties

    The properties file defines the logging configuration:

    # Define the logger
    kafka.root.logger.level="INFO"
    # ...
  2. Define external logging in the spec of the resource, setting the logging.valueFrom.configMapKeyRef.name to the name of the ConfigMap and logging.valueFrom.configMapKeyRef.key to the key in this ConfigMap.

    spec:
      # ...
      logging:
        type: external
        valueFrom:
          configMapKeyRef:
            name: logging-configmap
            key: log4j.properties
  3. Create or update the resource.

    kubectl apply -f KAFKA-CONFIG-FILE

2.8.2. Adding logging filters to Operators

If you are using a ConfigMap to configure the (log4j2) logging levels for Strimzi Operators, you can also define logging filters to limit what’s returned in the log.

Logging filters are useful when you have a large number of logging messages. Suppose you set the log level for the logger as DEBUG (rootLogger.level="DEBUG"). Logging filters reduce the number of logs returned for the logger at that level, so you can focus on a specific resource. When the filter is set, only log messages matching the filter are logged.

Filters use markers to specify what to include in the log. You specify a kind, namespace and name for the marker. For example, if a Kafka cluster is failing, you can isolate the logs by specifying the kind as Kafka, and use the namespace and name of the failing cluster.

This example shows a marker filter for a Kafka cluster named my-kafka-cluster.

Basic logging filter configuration
rootLogger.level="INFO"
appender.console.filter.filter1.type=MarkerFilter (1)
appender.console.filter.filter1.onMatch=ACCEPT (2)
appender.console.filter.filter1.onMismatch=DENY (3)
appender.console.filter.filter1.marker=Kafka(my-namespace/my-kafka-cluster) (4)
  1. The MarkerFilter type compares a specified marker for filtering.

  2. The onMatch property accepts the log if the marker matches.

  3. The onMismatch property rejects the log if the marker does not match.

  4. The marker used for filtering is in the format KIND(NAMESPACE/NAME-OF-RESOURCE).

You can create one or more filters. Here, the log is filtered for two Kafka clusters.

Multiple logging filter configuration
appender.console.filter.filter1.type=MarkerFilter
appender.console.filter.filter1.onMatch=ACCEPT
appender.console.filter.filter1.onMismatch=DENY
appender.console.filter.filter1.marker=Kafka(my-namespace/my-kafka-cluster-1)
appender.console.filter.filter2.type=MarkerFilter
appender.console.filter.filter2.onMatch=ACCEPT
appender.console.filter.filter2.onMismatch=DENY
appender.console.filter.filter2.marker=Kafka(my-namespace/my-kafka-cluster-2)
Adding filters to the Cluster Operator

To add filters to the Cluster Operator, update its logging ConfigMap YAML file (install/cluster-operator/050-ConfigMap-strimzi-cluster-operator.yaml).

Procedure
  1. Update the 050-ConfigMap-strimzi-cluster-operator.yaml file to add the filter properties to the ConfigMap.

    In this example, the filter properties return logs only for the my-kafka-cluster Kafka cluster:

    kind: ConfigMap
    apiVersion: v1
    metadata:
      name: strimzi-cluster-operator
    data:
      log4j2.properties:
        #...
        appender.console.filter.filter1.type=MarkerFilter
        appender.console.filter.filter1.onMatch=ACCEPT
        appender.console.filter.filter1.onMismatch=DENY
        appender.console.filter.filter1.marker=Kafka(my-namespace/my-kafka-cluster)

    Alternatively, edit the ConfigMap directly:

    kubectl edit configmap strimzi-cluster-operator
  2. If you updated the YAML file instead of editing the ConfigMap directly, apply the changes by deploying the ConfigMap:

    kubectl create -f install/cluster-operator/050-ConfigMap-strimzi-cluster-operator.yaml
Adding filters to the Topic Operator or User Operator

To add filters to the Topic Operator or User Operator, create or edit a logging ConfigMap.

In this procedure a logging ConfigMap is created with filters for the Topic Operator. The same approach is used for the User Operator.

Procedure
  1. Create the ConfigMap.

    You can create the ConfigMap as a YAML file or from a properties file.

    In this example, the filter properties return logs only for the my-topic topic:

    kind: ConfigMap
    apiVersion: v1
    metadata:
      name: logging-configmap
    data:
      log4j2.properties:
        rootLogger.level="INFO"
        appender.console.filter.filter1.type=MarkerFilter
        appender.console.filter.filter1.onMatch=ACCEPT
        appender.console.filter.filter1.onMismatch=DENY
        appender.console.filter.filter1.marker=KafkaTopic(my-namespace/my-topic)

    If you are using a properties file, specify the file at the command line:

    kubectl create configmap logging-configmap --from-file=log4j2.properties

    The properties file defines the logging configuration:

    # Define the logger
    rootLogger.level="INFO"
    # Set the filters
    appender.console.filter.filter1.type=MarkerFilter
    appender.console.filter.filter1.onMatch=ACCEPT
    appender.console.filter.filter1.onMismatch=DENY
    appender.console.filter.filter1.marker=KafkaTopic(my-namespace/my-topic)
    # ...
  2. Define external logging in the spec of the resource, setting the logging.valueFrom.configMapKeyRef.name to the name of the ConfigMap and logging.valueFrom.configMapKeyRef.key to the key in this ConfigMap.

    For the Topic Operator, logging is specified in the topicOperator configuration of the Kafka resource.

    spec:
      # ...
      entityOperator:
        topicOperator:
          logging:
            type: external
            valueFrom:
              configMapKeyRef:
                name: logging-configmap
                key: log4j2.properties
  3. Apply the changes by deploying the Cluster Operator:

create -f install/cluster-operator -n my-cluster-operator-namespace

2.9. Loading configuration values from external sources

Use the Kubernetes Configuration Provider plugin to load configuration data from external sources. Load data from Kubernetes Secrets or ConfigMaps.

Suppose you have a Secret that’s managed outside the Kafka namespace, or outside the Kafka cluster. The provider allows you to reference the values of the Secret in your configuration without extracting the files. You just need to tell the provider what Secret to use and provide access rights. The provider loads the data without needing to restart the Kafka component, even when using a new Secret or ConfigMap. This capability avoids disruption when a Kafka Connect instance hosts multiple connectors.

The provider operates independently of Strimzi. You can use it to load configuration data for all Kafka components, including producers and consumers. Use it, for example, to provide the credentials for Kafka Connect connector configuration.

In this procedure, an external ConfigMap provides configuration properties for a connector.

Note
Kubernetes Configuration Provider can’t use mounted files. For example, it can’t load values that need the location of a truststore or keystore. Instead, you can mount ConfigMaps or Secrets into a Kafka Connect pod as environment variables or volumes. You add configuration using the externalConfiguration property in KafkaConnect.spec. You don’t need to set up access rights with this approach. However, Kafka Connect will need a restart when using a new Secret or ConfigMap for a connector. This will cause disruption to all the Kafka Connect instance’s connectors.
Prerequisites
  • A Kubernetes cluster is available.

  • A Kafka cluster is running.

  • The Cluster Operator is running.

Procedure
  1. Create a ConfigMap or Secret that contains the configuration properties.

    In this example, a ConfigMap named my-connector-configuration contains connector properties:

    Example ConfigMap with connector properties
    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: my-connector-configuration
    data:
      option1: value1
      option2: value2
  2. Specify the Kubernetes Configuration Provider in the Kafka Connect configuration.

    The specification shown here can support loading values from Secrets and ConfigMaps.

    Example external volumes set to values from a ConfigMap
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaConnect
    metadata:
      name: my-connect
      annotations:
        strimzi.io/use-connector-resources: "true"
    spec:
      # ...
      config:
        # ...
        config.providers: secrets,configmaps (1)
        config.providers.secrets.class: io.strimzi.kafka.KubernetesSecretConfigProvider (2)
        config.providers.configmaps.class: io.strimzi.kafka.KubernetesConfigMapConfigProvider (3)
      # ...
    1. The alias for the configuration provider is used to define other configuration parameters. The provider parameters use the alias from config.providers, taking the form config.providers.${alias}.class.

    2. KubernetesSecretConfigProvider provides values from Secrets.

    3. KubernetesConfigMapConfigProvider provides values from ConfigMaps.

  3. Create or update the resource to enable the provider.

    kubectl apply -f KAFKA-CONNECT-CONFIG-FILE
  4. Create a role that permits access to the values in the external ConfigMap.

    Example role to access values from a ConfigMap
    apiVersion: rbac.authorization.k8s.io/v1
    kind: Role
    metadata:
      name: connector-configuration-role
    rules:
    - apiGroups: [""]
      resources: ["configmaps"]
      resourceNames: ["my-connector-configuration"]
      verbs: ["get"]
    # ...

    The rule gives the role permission to access the my-connector-configuration ConfigMap.

  5. Create a role binding to permit access to the namespace that contains the ConfigMap.

    Example role binding to access the namespace that contains the ConfigMap
    apiVersion: rbac.authorization.k8s.io/v1
    kind: RoleBinding
    metadata:
      name: connector-configuration-role-binding
    subjects:
    - kind: ServiceAccount
      name: my-connect-connect
      namespace: my-project
    roleRef:
      kind: Role
      name: connector-configuration-role
      apiGroup: rbac.authorization.k8s.io
    # ...

    The role binding gives the role permission to access the my-project namespace.

    The service account must be the same one used by the Kafka Connect deployment. The service account name format is CLUSTER_NAME-connect, where CLUSTER_NAME is the name of the KafkaConnect custom resource.

  6. Reference the ConfigMap in the connector configuration.

    Example connector configuration referencing the ConfigMap
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaConnector
    metadata:
      name: my-connector
      labels:
        strimzi.io/cluster: my-connect
    spec:
      # ...
      config:
        option: ${configmaps:my-project/my-connector-configuration:option1}
        # ...
    # ...

    Placeholders for the property values in the ConfigMap are referenced in the connector configuration. The placeholder structure is configmaps:PATH-AND-FILE-NAME:PROPERTY. KubernetesConfigMapConfigProvider reads and extracts the option1 property value from the external ConfigMap.

3. Configuring external listeners

Use an external listener to expose your Strimzi Kafka cluster to a client outside a Kubernetes environment.

Specify the connection type to expose Kafka in the external listener configuration.

  • nodeport uses NodePort type Services

  • loadbalancer uses Loadbalancer type Services

  • ingress uses Kubernetes Ingress and the NGINX Ingress Controller for Kubernetes

  • route uses OpenShift Routes and the HAProxy router

For more information on listener configuration, see GenericKafkaListener schema reference.

Note
route is only supported on OpenShift
Additional resources

3.1. Accessing Kafka using node ports

This procedure describes how to access a Strimzi Kafka cluster from an external client using node ports.

To connect to a broker, you need a hostname and port number for the Kafka bootstrap address, as well as the certificate used for authentication.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Configure a Kafka resource with an external listener set to the nodeport type.

    For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        listeners:
          - name: external
            port: 9094
            type: nodeport
            tls: true
            authentication:
              type: tls
            # ...
        # ...
      zookeeper:
        # ...
  2. Create or update the resource.

    kubectl apply -f KAFKA-CONFIG-FILE

    NodePort type services are created for each Kafka broker, as well as an external bootstrap service. The bootstrap service routes external traffic to the Kafka brokers. Node addresses used for connection are propagated to the status of the Kafka custom resource.

    The cluster CA certificate to verify the identity of the kafka brokers is also created with the same name as the Kafka resource.

  3. Retrieve the bootstrap address you can use to access the Kafka cluster from the status of the Kafka resource.

    kubectl get kafka KAFKA-CLUSTER-NAME -o=jsonpath='{.status.listeners[?(@.type=="external")].bootstrapServers}{"\n"}'
  4. If TLS encryption is enabled, extract the public certificate of the broker certification authority.

    kubectl get secret KAFKA-CLUSTER-NAME-cluster-ca-cert -o jsonpath='{.data.ca\.crt}' | base64 -d > ca.crt

    Use the extracted certificate in your Kafka client to configure TLS connection. If you enabled any authentication, you will also need to configure SASL or TLS authentication.

3.2. Accessing Kafka using loadbalancers

This procedure describes how to access a Strimzi Kafka cluster from an external client using loadbalancers.

To connect to a broker, you need the address of the bootstrap loadbalancer, as well as the certificate used for TLS encryption.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Configure a Kafka resource with an external listener set to the loadbalancer type.

    For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        listeners:
          - name: external
            port: 9094
            type: loadbalancer
            tls: true
            # ...
        # ...
      zookeeper:
        # ...
  2. Create or update the resource.

    kubectl apply -f KAFKA-CONFIG-FILE

    loadbalancer type services and loadbalancers are created for each Kafka broker, as well as an external bootstrap service. The bootstrap service routes external traffic to all Kafka brokers. DNS names and IP addresses used for connection are propagated to the status of each service.

    The cluster CA certificate to verify the identity of the kafka brokers is also created with the same name as the Kafka resource.

  3. Retrieve the address of the bootstrap service you can use to access the Kafka cluster from the status of the Kafka resource.

    kubectl get kafka KAFKA-CLUSTER-NAME -o=jsonpath='{.status.listeners[?(@.type=="external")].bootstrapServers}{"\n"}'
  4. If TLS encryption is enabled, extract the public certificate of the broker certification authority.

    kubectl get secret KAFKA-CLUSTER-NAME-cluster-ca-cert -o jsonpath='{.data.ca\.crt}' | base64 -d > ca.crt

    Use the extracted certificate in your Kafka client to configure TLS connection. If you enabled any authentication, you will also need to configure SASL or TLS authentication.

3.3. Accessing Kafka using ingress

This procedure shows how to access a Strimzi Kafka cluster from an external client outside of Kubernetes using Nginx Ingress.

To connect to a broker, you need a hostname (advertised address) for the Ingress bootstrap address, as well as the certificate used for authentication.

For access using Ingress, the port is always 443.

TLS passthrough

Kafka uses a binary protocol over TCP, but the NGINX Ingress Controller for Kubernetes is designed to work with the HTTP protocol. To be able to pass the Kafka connections through the Ingress, Strimzi uses the TLS passthrough feature of the NGINX Ingress Controller for Kubernetes. Ensure TLS passthrough is enabled in your NGINX Ingress Controller for Kubernetes deployment.

Because it is using the TLS passthrough functionality, TLS encryption cannot be disabled when exposing Kafka using Ingress.

For more information about enabling TLS passthrough, see TLS passthrough documentation.

Prerequisites
Procedure
  1. Configure a Kafka resource with an external listener set to the ingress type.

    Specify the Ingress hosts for the bootstrap service and Kafka brokers.

    For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        listeners:
          - name: external
            port: 9094
            type: ingress
            tls: true
            authentication:
              type: tls
            configuration: (1)
              bootstrap:
                host: bootstrap.myingress.com
              brokers:
              - broker: 0
                host: broker-0.myingress.com
              - broker: 1
                host: broker-1.myingress.com
              - broker: 2
                host: broker-2.myingress.com
        # ...
      zookeeper:
        # ...
    1. Ingress hosts for the bootstrap service and Kafka brokers.

  2. Create or update the resource.

    kubectl apply -f KAFKA-CONFIG-FILE

    ClusterIP type services are created for each Kafka broker, as well as an additional bootstrap service. These services are used by the Ingress controller to route traffic to the Kafka brokers. An Ingress resource is also created for each service to expose them using the Ingress controller. The Ingress hosts are propagated to the status of each service.

    The cluster CA certificate to verify the identity of the kafka brokers is also created with the same name as the Kafka resource.

    Use the address for the bootstrap host you specified in the configuration and port 443 (BOOTSTRAP-HOST:443) in your Kafka client as the bootstrap address to connect to the Kafka cluster.

  3. Extract the public certificate of the broker certificate authority.

    kubectl get secret KAFKA-CLUSTER-NAME-cluster-ca-cert -o jsonpath='{.data.ca\.crt}' | base64 -d > ca.crt

    Use the extracted certificate in your Kafka client to configure the TLS connection. If you enabled any authentication, you will also need to configure SASL or TLS authentication.

3.4. Accessing Kafka using OpenShift routes

This procedure describes how to access a Strimzi Kafka cluster from an external client outside of OpenShift using routes.

To connect to a broker, you need a hostname for the route bootstrap address, as well as the certificate used for TLS encryption.

For access using routes, the port is always 443.

Prerequisites
  • An OpenShift cluster

  • A running Cluster Operator

Procedure
  1. Configure a Kafka resource with an external listener set to the route type.

    For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    metadata:
      labels:
        app: my-cluster
      name: my-cluster
      namespace: myproject
    spec:
      kafka:
        # ...
        listeners:
          - name: listener1
            port: 9094
            type: route
            tls: true
            # ...
        # ...
      zookeeper:
        # ...
    Warning
    An OpenShift Route address comprises the name of the Kafka cluster, the name of the listener, and the name of the namespace it is created in. For example, my-cluster-kafka-listener1-bootstrap-myproject (CLUSTER-NAME-kafka-LISTENER-NAME-bootstrap-NAMESPACE). Be careful that the whole length of the address does not exceed a maximum limit of 63 characters.
  2. Create or update the resource.

    kubectl apply -f KAFKA-CONFIG-FILE

    ClusterIP type services are created for each Kafka broker, as well as an external bootstrap service. The services route the traffic from the OpenShift Routes to the Kafka brokers. An OpenShift Route resource is also created for each service to expose them using the HAProxy load balancer. DNS addresses used for connection are propagated to the status of each service.

    The cluster CA certificate to verify the identity of the kafka brokers is also created with the same name as the Kafka resource.

  3. Retrieve the address of the bootstrap service you can use to access the Kafka cluster from the status of the Kafka resource.

    kubectl get kafka KAFKA-CLUSTER-NAME -o=jsonpath='{.status.listeners[?(@.type=="external")].bootstrapServers}{"\n"}'
  4. Extract the public certificate of the broker certification authority.

    kubectl get secret KAFKA-CLUSTER-NAME-cluster-ca-cert -o jsonpath='{.data.ca\.crt}' | base64 -d > ca.crt

    Use the extracted certificate in your Kafka client to configure TLS connection. If you enabled any authentication, you will also need to configure SASL or TLS authentication.

4. Managing secure access to Kafka

You can secure your Kafka cluster by managing the access each client has to the Kafka brokers.

A secure connection between Kafka brokers and clients can encompass:

  • Encryption for data exchange

  • Authentication to prove identity

  • Authorization to allow or decline actions executed by users

This chapter explains how to set up secure connections between Kafka brokers and clients, with sections describing:

  • Security options for Kafka clusters and clients

  • How to secure Kafka brokers

  • How to use an authorization server for OAuth 2.0 token-based authentication and authorization

4.1. Security options for Kafka

Use the Kafka resource to configure the mechanisms used for Kafka authentication and authorization.

4.1.1. Listener authentication

For clients inside the Kubernetes cluster, you can create plain (without encryption) or tls internal listeners.

For clients outside the Kubernetes cluster, you create external listeners and specify a connection mechanism, which can be nodeport, loadbalancer, ingress or route (on OpenShift).

For more information on the configuration options for connecting an external client, see Configuring external listeners.

Supported authentication options:

  1. Mutual TLS authentication (only on the listeners with TLS enabled encryption)

  2. SCRAM-SHA-512 authentication

  3. OAuth 2.0 token based authentication

The authentication option you choose depends on how you wish to authenticate client access to Kafka brokers.

options for listener authentication configuration
Figure 3. Kafka listener authentication options

The listener authentication property is used to specify an authentication mechanism specific to that listener.

If no authentication property is specified then the listener does not authenticate clients which connect through that listener. The listener will accept all connections without authentication.

Authentication must be configured when using the User Operator to manage KafkaUsers.

The following example shows:

  • A plain listener configured for SCRAM-SHA-512 authentication

  • A tls listener with mutual TLS authentication

  • An external listener with mutual TLS authentication

Each listener is configured with a unique name and port within a Kafka cluster.

Note
Listeners cannot be configured to use the ports reserved for inter-broker communication (9091 or 9090) and metrics (9404).
An example showing listener authentication configuration
# ...
listeners:
  - name: plain
    port: 9092
    type: internal
    tls: true
    authentication:
      type: scram-sha-512
  - name: tls
    port: 9093
    type: internal
    tls: true
    authentication:
      type: tls
  - name: external
    port: 9094
    type: loadbalancer
    tls: true
    authentication:
      type: tls
# ...
Mutual TLS authentication

Mutual TLS authentication is always used for the communication between Kafka brokers and ZooKeeper pods.

Strimzi can configure Kafka to use TLS (Transport Layer Security) to provide encrypted communication between Kafka brokers and clients either with or without mutual authentication. For mutual, or two-way, authentication, both the server and the client present certificates. When you configure mutual authentication, the broker authenticates the client (client authentication) and the client authenticates the broker (server authentication).

Note
TLS authentication is more commonly one-way, with one party authenticating the identity of another. For example, when HTTPS is used between a web browser and a web server, the browser obtains proof of the identity of the web server.
SCRAM-SHA-512 authentication

SCRAM (Salted Challenge Response Authentication Mechanism) is an authentication protocol that can establish mutual authentication using passwords. Strimzi can configure Kafka to use SASL (Simple Authentication and Security Layer) SCRAM-SHA-512 to provide authentication on both unencrypted and encrypted client connections.

When SCRAM-SHA-512 authentication is used with a TLS client connection, the TLS protocol provides the encryption, but is not used for authentication.

The following properties of SCRAM make it safe to use SCRAM-SHA-512 even on unencrypted connections:

  • The passwords are not sent in the clear over the communication channel. Instead the client and the server are each challenged by the other to offer proof that they know the password of the authenticating user.

  • The server and client each generate a new challenge for each authentication exchange. This means that the exchange is resilient against replay attacks.

When a KafkaUser.spec.authentication.type is configured with scram-sha-512 the User Operator will generate a random 12-character password consisting of upper and lowercase ASCII letters and numbers.

Network policies

Strimzi automatically creates a NetworkPolicy resource for every listener that is enabled on a Kafka broker. By default, a NetworkPolicy grants access to a listener to all applications and namespaces.

If you want to restrict access to a listener at the network level to only selected applications or namespaces, use the networkPolicyPeers property.

Use network policies as part of the listener authentication configuration. Each listener can have a different networkPolicyPeers configuration.

For more information, refer to the Listener network policies section and the NetworkPolicyPeer API reference.

Note
Your configuration of Kubernetes must support ingress NetworkPolicies in order to use network policies in Strimzi.
Additional listener configuration options

You can use the properties of the GenericKafkaListenerConfiguration schema to add further configuration to listeners.

4.1.2. Kafka authorization

You can configure authorization for Kafka brokers using the authorization property in the Kafka.spec.kafka resource. If the authorization property is missing, no authorization is enabled and clients have no restrictions. When enabled, authorization is applied to all enabled listeners. The authorization method is defined in the type field.

Supported authorization options:

options for kafks authorization configuration
Figure 4. Kafka cluster authorization options
Super users

Super users can access all resources in your Kafka cluster regardless of any access restrictions, and are supported by all authorization mechanisms.

To designate super users for a Kafka cluster, add a list of user principals to the superUsers property. If a user uses TLS client authentication, their username is the common name from their certificate subject prefixed with CN=.

An example configuration with super users
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
  namespace: myproject
spec:
  kafka:
    # ...
    authorization:
      type: simple
      superUsers:
        - CN=client_1
        - user_2
        - CN=client_3
    # ...

4.2. Security options for Kafka clients

Use the KafkaUser resource to configure the authentication mechanism, authorization mechanism, and access rights for Kafka clients. In terms of configuring security, clients are represented as users.

You can authenticate and authorize user access to Kafka brokers. Authentication permits access, and authorization constrains the access to permissible actions.

You can also create super users that have unconstrained access to Kafka brokers.

The authentication and authorization mechanisms must match the specification for the listener used to access the Kafka brokers.

4.2.1. Identifying a Kafka cluster for user handling

A KafkaUser resource includes a label that defines the appropriate name of the Kafka cluster (derived from the name of the Kafka resource) to which it belongs.

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaUser
metadata:
  name: my-user
  labels:
    strimzi.io/cluster: my-cluster

The label is used by the User Operator to identify the KafkaUser resource and create a new user, and also in subsequent handling of the user.

If the label does not match the Kafka cluster, the User Operator cannot identify the KafkaUser and the user is not created.

If the status of the KafkaUser resource remains empty, check your label.

4.2.2. User authentication

User authentication is configured using the authentication property in KafkaUser.spec. The authentication mechanism enabled for the user is specified using the type field.

Supported authentication mechanisms:

  • TLS client authentication

  • SCRAM-SHA-512 authentication

When no authentication mechanism is specified, the User Operator does not create the user or its credentials.

TLS Client Authentication

To use TLS client authentication, you set the type field to tls.

An example KafkaUser with TLS client authentication enabled
apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaUser
metadata:
  name: my-user
  labels:
    strimzi.io/cluster: my-cluster
spec:
  authentication:
    type: tls
  # ...

When the user is created by the User Operator, it creates a new Secret with the same name as the KafkaUser resource. The Secret contains a private and public key for TLS client authentication. The public key is contained in a user certificate, which is signed by the client Certificate Authority (CA).

All keys are in X.509 format.

Secrets provide private keys and certificates in PEM and PKCS #12 formats.

For more information on securing Kafka communication with Secrets, see Managing TLS certificates.

An example Secret with user credentials
apiVersion: v1
kind: Secret
metadata:
  name: my-user
  labels:
    strimzi.io/kind: KafkaUser
    strimzi.io/cluster: my-cluster
type: Opaque
data:
  ca.crt: # Public key of the client CA
  user.crt: # User certificate that contains the public key of the user
  user.key: # Private key of the user
  user.p12: # PKCS #12 archive file for storing certificates and keys
  user.password: # Password for protecting the PKCS #12 archive file
SCRAM-SHA-512 Authentication

To use the SCRAM-SHA-512 authentication mechanism, you set the type field to scram-sha-512.

An example KafkaUser with SCRAM-SHA-512 authentication enabled
apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaUser
metadata:
  name: my-user
  labels:
    strimzi.io/cluster: my-cluster
spec:
  authentication:
    type: scram-sha-512
  # ...

When the user is created by the User Operator, it creates a new secret with the same name as the KafkaUser resource. The secret contains the generated password in the password key, which is encoded with base64. In order to use the password, it must be decoded.

An example Secret with user credentials
apiVersion: v1
kind: Secret
metadata:
  name: my-user
  labels:
    strimzi.io/kind: KafkaUser
    strimzi.io/cluster: my-cluster
type: Opaque
data:
  password: Z2VuZXJhdGVkcGFzc3dvcmQ= (1)
  sasl.jaas.config: b3JnLmFwYWNoZS5rYWZrYS5jb21tb24uc2VjdXJpdHkuc2NyYW0uU2NyYW1Mb2dpbk1vZHVsZSByZXF1aXJlZCB1c2VybmFtZT0ibXktdXNlciIgcGFzc3dvcmQ9ImdlbmVyYXRlZHBhc3N3b3JkIjsK (2)
  1. The generated password, base64 encoded.

  2. The JAAS configuration string for SASL SCRAM-SHA-512 authentication, base64 encoded.

Decoding the generated password:

echo "Z2VuZXJhdGVkcGFzc3dvcmQ=" | base64 --decode

4.2.3. User authorization

User authorization is configured using the authorization property in KafkaUser.spec. The authorization type enabled for a user is specified using the type field.

To use simple authorization, you set the type property to simple in KafkaUser.spec.authorization. Simple authorization uses the default Kafka authorization plugin, AclAuthorizer.

Alternatively, you can use OPA authorization, or if you are already using OAuth 2.0 token based authentication, you can also use OAuth 2.0 authorization.

If no authorization is specified, the User Operator does not provision any access rights for the user. Whether such a KafkaUser can still access resources depends on the authorizer being used. For example, for the AclAuthorizer this is determined by its allow.everyone.if.no.acl.found configuration.

ACL rules

AclAuthorizer uses ACL rules to manage access to Kafka brokers.

ACL rules grant access rights to the user, which you specify in the acls property.

For more information about the AclRule object, see the AclRule schema reference.

Super user access to Kafka brokers

If a user is added to a list of super users in a Kafka broker configuration, the user is allowed unlimited access to the cluster regardless of any authorization constraints defined in ACLs in KafkaUser.

For more information on configuring super user access to brokers, see Kafka authorization.

User quotas

You can configure the spec for the KafkaUser resource to enforce quotas so that a user does not exceed access to Kafka brokers based on a byte threshold or a time limit of CPU utilization.

An example KafkaUser with user quotas
apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaUser
metadata:
  name: my-user
  labels:
    strimzi.io/cluster: my-cluster
spec:
  # ...
  quotas:
    producerByteRate: 1048576 (1)
    consumerByteRate: 2097152 (2)
    requestPercentage: 55 (3)
  1. Byte-per-second quota on the amount of data the user can push to a Kafka broker

  2. Byte-per-second quota on the amount of data the user can fetch from a Kafka broker

  3. CPU utilization limit as a percentage of time for a client group

For more information on these properties, see the KafkaUserQuotas schema reference.

4.3. Securing access to Kafka brokers

To establish secure access to Kafka brokers, you configure and apply:

  • A Kafka resource to:

    • Create listeners with a specified authentication type

    • Configure authorization for the whole Kafka cluster

  • A KafkaUser resource to access the Kafka brokers securely through the listeners

Configure the Kafka resource to set up:

  • Listener authentication

  • Network policies that restrict access to Kafka listeners

  • Kafka authorization

  • Super users for unconstrained access to brokers

Authentication is configured independently for each listener. Authorization is always configured for the whole Kafka cluster.

The Cluster Operator creates the listeners and sets up the cluster and client certificate authority (CA) certificates to enable authentication within the Kafka cluster.

You can replace the certificates generated by the Cluster Operator by installing your own certificates. You can also configure your listener to use a Kafka listener certificate managed by an external Certificate Authority. Certificates are available in PKCS #12 format (.p12) and PEM (.crt) formats.

Use KafkaUser to enable the authentication and authorization mechanisms that a specific client uses to access Kafka.

Configure the KafkaUser resource to set up:

  • Authentication to match the enabled listener authentication

  • Authorization to match the enabled Kafka authorization

  • Quotas to control the use of resources by clients

The User Operator creates the user representing the client and the security credentials used for client authentication, based on the chosen authentication type.

Additional resources

For more information about the schema for:

4.3.1. Securing Kafka brokers

This procedure shows the steps involved in securing Kafka brokers when running Strimzi.

The security implemented for Kafka brokers must be compatible with the security implemented for the clients requiring access.

  • Kafka.spec.kafka.listeners[*].authentication matches KafkaUser.spec.authentication

  • Kafka.spec.kafka.authorization matches KafkaUser.spec.authorization

The steps show the configuration for simple authorization and a listener using TLS authentication. For more information on listener configuration, see GenericKafkaListener schema reference.

Alternatively, you can use SCRAM-SHA or OAuth 2.0 for listener authentication, and OAuth 2.0 or OPA for Kafka authorization.

Procedure
  1. Configure the Kafka resource.

    1. Configure the authorization property for authorization.

    2. Configure the listeners property to create a listener with authentication.

      For example:

      apiVersion: kafka.strimzi.io/v1beta2
      kind: Kafka
      spec:
        kafka:
          # ...
          authorization: (1)
            type: simple
            superUsers: (2)
              - CN=client_1
              - user_2
              - CN=client_3
          listeners:
            - name: tls
              port: 9093
              type: internal
              tls: true
              authentication:
                type: tls (3)
          # ...
        zookeeper:
          # ...
      1. Authorization enables simple authorization on the Kafka broker using the AclAuthorizer Kafka plugin.

      2. List of user principals with unlimited access to Kafka. CN is the common name from the client certificate when TLS authentication is used.

      3. Listener authentication mechanisms may be configured for each listener, and specified as mutual TLS, SCRAM-SHA-512 or token-based OAuth 2.0.

      If you are configuring an external listener, the configuration is dependent on the chosen connection mechanism.

  2. Create or update the Kafka resource.

    kubectl apply -f KAFKA-CONFIG-FILE

    The Kafka cluster is configured with a Kafka broker listener using TLS authentication.

    A service is created for each Kafka broker pod.

    A service is created to serve as the bootstrap address for connection to the Kafka cluster.

    The cluster CA certificate to verify the identity of the kafka brokers is also created with the same name as the Kafka resource.

4.3.2. Securing user access to Kafka

Use the properties of the KafkaUser resource to configure a Kafka user.

You can use kubectl apply to create or modify users, and kubectl delete to delete existing users.

For example:

  • kubectl apply -f USER-CONFIG-FILE

  • kubectl delete KafkaUser USER-NAME

When you configure the KafkaUser authentication and authorization mechanisms, ensure they match the equivalent Kafka configuration:

  • KafkaUser.spec.authentication matches Kafka.spec.kafka.listeners[*].authentication

  • KafkaUser.spec.authorization matches Kafka.spec.kafka.authorization

This procedure shows how a user is created with TLS authentication. You can also create a user with SCRAM-SHA authentication.

The authentication required depends on the type of authentication configured for the Kafka broker listener.

Note
Authentication between Kafka users and Kafka brokers depends on the authentication settings for each. For example, it is not possible to authenticate a user with TLS if it is not also enabled in the Kafka configuration.
Prerequisites

The authentication type in KafkaUser should match the authentication configured in Kafka brokers.

Procedure
  1. Configure the KafkaUser resource.

    For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaUser
    metadata:
      name: my-user
      labels:
        strimzi.io/cluster: my-cluster
    spec:
      authentication: (1)
        type: tls
      authorization:
        type: simple (2)
        acls:
          - resource:
              type: topic
              name: my-topic
              patternType: literal
            operation: Read
          - resource:
              type: topic
              name: my-topic
              patternType: literal
            operation: Describe
          - resource:
              type: group
              name: my-group
              patternType: literal
            operation: Read
    1. User authentication mechanism, defined as mutual tls or scram-sha-512.

    2. Simple authorization, which requires an accompanying list of ACL rules.

  2. Create or update the KafkaUser resource.

    kubectl apply -f USER-CONFIG-FILE

    The user is created, as well as a Secret with the same name as the KafkaUser resource. The Secret contains a private and public key for TLS client authentication.

For information on configuring a Kafka client with properties for secure connection to Kafka brokers, see Setting up access for clients outside of Kubernetes in the Deploying and Upgrading Strimzi guide.

4.3.3. Restricting access to Kafka listeners using network policies

You can restrict access to a listener to only selected applications by using the networkPolicyPeers property.

Prerequisites
  • A Kubernetes cluster with support for Ingress NetworkPolicies.

  • The Cluster Operator is running.

Procedure
  1. Open the Kafka resource.

  2. In the networkPolicyPeers property, define the application pods or namespaces that will be allowed to access the Kafka cluster.

    For example, to configure a tls listener to allow connections only from application pods with the label app set to kafka-client:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      kafka:
        # ...
        listeners:
          - name: tls
            port: 9093
            type: internal
            tls: true
            authentication:
              type: tls
            networkPolicyPeers:
              - podSelector:
                  matchLabels:
                    app: kafka-client
        # ...
      zookeeper:
        # ...
  3. Create or update the resource.

    Use kubectl apply:

    kubectl apply -f your-file
Additional resources

4.4. Using OAuth 2.0 token-based authentication

Strimzi supports the use of OAuth 2.0 authentication using the OAUTHBEARER and PLAIN mechanisms.

OAuth 2.0 enables standardized token-based authentication and authorization between applications, using a central authorization server to issue tokens that grant limited access to resources.

Kafka brokers and clients both need to be configured to use OAuth 2.0. You can configure OAuth 2.0 authentication, then OAuth 2.0 authorization.

Note

OAuth 2.0 authentication can be used in conjunction with Kafka authorization.

Using OAuth 2.0 authentication, application clients can access resources on application servers (called resource servers) without exposing account credentials.

The application client passes an access token as a means of authenticating, which application servers can also use to determine the level of access to grant. The authorization server handles the granting of access and inquiries about access.

In the context of Strimzi:

  • Kafka brokers act as OAuth 2.0 resource servers

  • Kafka clients act as OAuth 2.0 application clients

Kafka clients authenticate to Kafka brokers. The brokers and clients communicate with the OAuth 2.0 authorization server, as necessary, to obtain or validate access tokens.

For a deployment of Strimzi, OAuth 2.0 integration provides:

  • Server-side OAuth 2.0 support for Kafka brokers

  • Client-side OAuth 2.0 support for Kafka MirrorMaker, Kafka Connect and the Kafka Bridge

Additional resources

4.4.1. OAuth 2.0 authentication mechanisms

Strimzi supports the OAUTHBEARER and PLAIN mechanisms for OAuth 2.0 authentication. Both mechanisms allow Kafka clients to establish authenticated sessions with Kafka brokers. The authentication flow between clients, the authorization server, and Kafka brokers is different for each mechanism.

We recommend that you configure clients to use OAUTHBEARER whenever possible. OAUTHBEARER provides a higher level of security than PLAIN because client credentials are never shared with Kafka brokers. Consider using PLAIN only with Kafka clients that do not support OAUTHBEARER.

If necessary, OAUTHBEARER and PLAIN can be enabled together, on the same oauth listener.

OAUTHBEARER overview

Kafka supports the OAUTHBEARER authentication mechanism, however it must be explicitly configured. Also, many Kafka client tools use libraries that provide basic support for OAUTHBEARER at the protocol level.

To ease application development, Strimzi provides an OAuth callback handler for the upstream Kafka Client Java libraries (but not for other libraries). Therefore, you do not need to write your own callback handlers for such clients. An application client can use the callback handler to provide the access token. Clients written in other languages, such as Go, must use custom code to connect to the authorization server and obtain the access token.

With OAUTHBEARER, the client initiates a session with the Kafka broker for credentials exchange, where credentials take the form of a bearer token provided by the callback handler. Using the callbacks, you can configure token provision in one of three ways:

  • Client ID and Secret (by using the OAuth 2.0 client credentials mechanism)

  • A long-lived access token, obtained manually at configuration time

  • A long-lived refresh token, obtained manually at configuration time

OAUTHBEARER is automatically enabled in the oauth listener configuration for the Kafka broker. You can set the enableOauthBearer property to true, though this is not required.

  # ...
  authentication:
    type: oauth
    # ...
    enableOauthBearer: true
Note

OAUTHBEARER authentication can only be used by Kafka clients that support the OAUTHBEARER mechanism at the protocol level.

PLAIN overview

PLAIN is a simple authentication mechanism used by all Kafka client tools (including developer tools such as kafkacat). To enable PLAIN to be used together with OAuth 2.0 authentication, Strimzi includes server-side callbacks and calls this OAuth 2.0 over PLAIN.

With the Strimzi implementation of PLAIN, the client credentials are not stored in ZooKeeper. Instead, client credentials are handled centrally behind a compliant authorization server, similar to when OAUTHBEARER authentication is used.

When used with the OAuth 2.0 over PLAIN callbacks, Kafka clients authenticate with Kafka brokers using either of the following methods:

  • Client ID and secret (by using the OAuth 2.0 client credentials mechanism)

  • A long-lived access token, obtained manually at configuration time

The client must be enabled to use PLAIN authentication, and provide a username and password. If the password is prefixed with $accessToken: followed by the value of the access token, the Kafka broker will interpret the password as the access token. Otherwise, the Kafka broker will interpret the username as the client ID and the password as the client secret.

If the password is set as the access token, the username must be set to the same principal name that the Kafka broker obtains from the access token. The process depends on how you configure username extraction using userNameClaim, fallbackUserNameClaim, fallbackUsernamePrefix, or userInfoEndpointUri. It also depends on your authorization server; in particular, how it maps client IDs to account names.

To use PLAIN, you must enable it in the oauth listener configuration for the Kafka broker.

In the following example, PLAIN is enabled in addition to OAUTHBEARER, which is enabled by default. If you want to use PLAIN only, you can disable OAUTHBEARER by setting enableOauthBearer to false.

  # ...
  authentication:
    type: oauth
    # ...
    enablePlain: true
    tokenEndpointUri: https://OAUTH-SERVER-ADDRESS/auth/realms/external/protocol/openid-connect/token

4.4.2. OAuth 2.0 Kafka broker configuration

Kafka broker configuration for OAuth 2.0 involves:

  • Creating the OAuth 2.0 client in the authorization server

  • Configuring OAuth 2.0 authentication in the Kafka custom resource

Note
In relation to the authorization server, Kafka brokers and Kafka clients are both regarded as OAuth 2.0 clients.
OAuth 2.0 client configuration on an authorization server

To configure a Kafka broker to validate the token received during session initiation, the recommended approach is to create an OAuth 2.0 client definition in an authorization server, configured as confidential, with the following client credentials enabled:

  • Client ID of kafka (for example)

  • Client ID and Secret as the authentication mechanism

Note
You only need to use a client ID and secret when using a non-public introspection endpoint of the authorization server. The credentials are not typically required when using public authorization server endpoints, as with fast local JWT token validation.
OAuth 2.0 authentication configuration in the Kafka cluster

To use OAuth 2.0 authentication in the Kafka cluster, you specify, for example, a TLS listener configuration for your Kafka cluster custom resource with the authentication method oauth:

Assigining the authentication method type for OAuth 2.0
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
spec:
  kafka:
    # ...
    listeners:
      - name: tls
        port: 9093
        type: internal
        tls: true
        authentication:
          type: oauth
      #...

You can configure plain, tls and external listeners, but it is recommended not to use plain listeners or external listeners with disabled TLS encryption with OAuth 2.0 as this creates a vulnerability to network eavesdropping and unauthorized access through token theft.

You configure an external listener with type: oauth for a secure transport layer to communicate with the client.

Using OAuth 2.0 with an external listener
# ...
listeners:
  - name: external
    port: 9094
    type: loadbalancer
    tls: true
    authentication:
      type: oauth
    #...

The tls property is false by default, so it must be enabled.

When you have defined the type of authentication as OAuth 2.0, you add configuration based on the type of validation, either as fast local JWT validation or token validation using an introspection endpoint.

The procedure to configure OAuth 2.0 for listeners, with descriptions and examples, is described in Configuring OAuth 2.0 support for Kafka brokers.

Fast local JWT token validation configuration

Fast local JWT token validation checks a JWT token signature locally.

The local check ensures that a token:

  • Conforms to type by containing a (typ) claim value of Bearer for an access token

  • Is valid (not expired)

  • Has an issuer that matches a validIssuerURI

You specify a validIssuerURI attribute when you configure the listener, so that any tokens not issued by the authorization server are rejected.

The authorization server does not need to be contacted during fast local JWT token validation. You activate fast local JWT token validation by specifying a jwksEndpointUri attribute, the endpoint exposed by the OAuth 2.0 authorization server. The endpoint contains the public keys used to validate signed JWT tokens, which are sent as credentials by Kafka clients.

Note
All communication with the authorization server should be performed using TLS encryption.

You can configure a certificate truststore as a Kubernetes Secret in your Strimzi project namespace, and use a tlsTrustedCertificates attribute to point to the Kubernetes Secret containing the truststore file.

You might want to configure a userNameClaim to properly extract a username from the JWT token. If you want to use Kafka ACL authorization, you need to identify the user by their username during authentication. (The sub claim in JWT tokens is typically a unique ID, not a username.)

Example configuration for fast local JWT token validation
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
spec:
  kafka:
    #...
    listeners:
      - name: tls
        port: 9093
        type: internal
        tls: true
        authentication:
          type: oauth
          validIssuerUri: <https://<auth-server-address>/auth/realms/tls>
          jwksEndpointUri: <https://<auth-server-address>/auth/realms/tls/protocol/openid-connect/certs>
          userNameClaim: preferred_username
          maxSecondsWithoutReauthentication: 3600
          tlsTrustedCertificates:
          - secretName: oauth-server-cert
            certificate: ca.crt
    #...
OAuth 2.0 introspection endpoint configuration

Token validation using an OAuth 2.0 introspection endpoint treats a received access token as opaque. The Kafka broker sends an access token to the introspection endpoint, which responds with the token information necessary for validation. Importantly, it returns up-to-date information if the specific access token is valid, and also information about when the token expires.

To configure OAuth 2.0 introspection-based validation, you specify an introspectionEndpointUri attribute rather than the jwksEndpointUri attribute specified for fast local JWT token validation. Depending on the authorization server, you typically have to specify a clientId and clientSecret, because the introspection endpoint is usually protected.

Example configuration for an introspection endpoint
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
spec:
  kafka:
    listeners:
      - name: tls
        port: 9093
        type: internal
        tls: true
        authentication:
          type: oauth
          clientId: kafka-broker
          clientSecret:
            secretName: my-cluster-oauth
            key: clientSecret
          validIssuerUri: <https://<auth-server-address>/auth/realms/tls>
          introspectionEndpointUri: <https://<auth-server-address>/auth/realms/tls/protocol/openid-connect/token/introspect>
          userNameClaim: preferred_username
          maxSecondsWithoutReauthentication: 3600
          tlsTrustedCertificates:
          - secretName: oauth-server-cert
            certificate: ca.crt

4.4.3. Session re-authentication for Kafka brokers

You can configure oauth listeners to use Kafka session re-authentication for OAuth 2.0 sessions between Kafka clients and Kafka brokers. This mechanism enforces the expiry of an authenticated session between the client and the broker after a defined period of time. When a session expires, the client immediately starts a new session by reusing the existing connection rather than dropping it.

Session re-authentication is disabled by default. To enable it, you set a time value for maxSecondsWithoutReauthentication in the oauth listener configuration. The same property is used to configure session re-authentication for OAUTHBEARER and PLAIN authentication. For an example configuration, see Configuring OAuth 2.0 support for Kafka brokers.

Session re-authentication must be supported by the Kafka client libraries used by the client.

Session re-authentication can be used with fast local JWT or introspection endpoint token validation.

Client re-authentication

When the broker’s authenticated session expires, the client must re-authenticate to the existing session by sending a new, valid access token to the broker, without dropping the connection.

If token validation is successful, a new client session is started using the existing connection. If the client fails to re-authenticate, the broker will close the connection if further attempts are made to send or receive messages. Java clients that use Kafka client library 2.2 or later automatically re-authenticate if the re-authentication mechanism is enabled on the broker.

Session re-authentication also applies to refresh tokens, if used. When the session expires, the client refreshes the access token by using its refresh token. The client then uses the new access token to re-authenticate to the existing session.

Session expiry for OAUTHBEARER and PLAIN

When session re-authentication is configured, session expiry works differently for OAUTHBEARER and PLAIN authentication.

For OAUTHBEARER and PLAIN, using the client ID and secret method:

  • The broker’s authenticated session will expire at the configured maxSecondsWithoutReauthentication.

  • The session will expire earlier if the access token expires before the configured time.

For PLAIN using the long-lived access token method:

  • The broker’s authenticated session will expire at the configured maxSecondsWithoutReauthentication.

  • Re-authentication will fail if the access token expires before the configured time. Although session re-authentication is attempted, PLAIN has no mechanism for refreshing tokens.

If maxSecondsWithoutReauthentication is not configured, OAUTHBEARER and PLAIN clients can remain connected to brokers indefinitely, without needing to re-authenticate. Authenticated sessions do not end with access token expiry. However, this can be considered when configuring authorization, for example, by using keycloak authorization or installing a custom authorizer.

4.4.4. OAuth 2.0 Kafka client configuration

A Kafka client is configured with either:

  • The credentials required to obtain a valid access token from an authorization server (client ID and Secret)

  • A valid long-lived access token or refresh token, obtained using tools provided by an authorization server

The only information ever sent to the Kafka broker is an access token. The credentials used to authenticate with the authorization server to obtain the access token are never sent to the broker.

When a client obtains an access token, no further communication with the authorization server is needed.

The simplest mechanism is authentication with a client ID and Secret. Using a long-lived access token, or a long-lived refresh token, adds more complexity because there is an additional dependency on authorization server tools.

Note
If you are using long-lived access tokens, you may need to configure the client in the authorization server to increase the maximum lifetime of the token.

If the Kafka client is not configured with an access token directly, the client exchanges credentials for an access token during Kafka session initiation by contacting the authorization server. The Kafka client exchanges either:

  • Client ID and Secret

  • Client ID, refresh token, and (optionally) a Secret

4.4.5. OAuth 2.0 client authentication flow

In this section, we explain and visualize the communication flow between Kafka client, Kafka broker, and authorization server during Kafka session initiation. The flow depends on the client and server configuration.

When a Kafka client sends an access token as credentials to a Kafka broker, the token needs to be validated.

Depending on the authorization server used, and the configuration options available, you may prefer to use:

  • Fast local token validation based on JWT signature checking and local token introspection, without contacting the authorization server

  • An OAuth 2.0 introspection endpoint provided by the authorization server

Using fast local token validation requires the authorization server to provide a JWKS endpoint with public certificates that are used to validate signatures on the tokens.

Another option is to use an OAuth 2.0 introspection endpoint on the authorization server. Each time a new Kafka broker connection is established, the broker passes the access token received from the client to the authorization server, and checks the response to confirm whether or not the token is valid.

Kafka client credentials can also be configured for:

  • Direct local access using a previously generated long-lived access token

  • Contact with the authorization server for a new access token to be issued

Note
An authorization server might only allow the use of opaque access tokens, which means that local token validation is not possible.
Example client authentication flows

Here you can see the communication flows, for different configurations of Kafka clients and brokers, during Kafka session authentication.

Client using client ID and secret, with broker delegating validation to authorization server

Client using client ID and secret with broker delegating validation to authorization server

  1. Kafka client requests access token from authorization server, using client ID and secret, and optionally a refresh token.

  2. Authorization server generates a new access token.

  3. Kafka client authenticates with the Kafka broker using the SASL OAUTHBEARER mechanism to pass the access token.

  4. Kafka broker validates the access token by calling a token introspection endpoint on authorization server, using its own client ID and secret.

  5. Kafka client session is established if the token is valid.

Client using client ID and secret, with broker performing fast local token validation

Client using client ID and secret with broker performing fast local token validation

  1. Kafka client authenticates with authorization server from the token endpoint, using a client ID and secret, and optionally a refresh token.

  2. Authorization server generates a new access token.

  3. Kafka client authenticates with the Kafka broker using the SASL OAUTHBEARER mechanism to pass the access token.

  4. Kafka broker validates the access token locally using a JWT token signature check, and local token introspection.

Client using long-lived access token, with broker delegating validation to authorization server

Client using long-lived access token with broker delegating validation to authorization server

  1. Kafka client authenticates with the Kafka broker using the SASL OAUTHBEARER mechanism to pass the long-lived access token.

  2. Kafka broker validates the access token by calling a token introspection endpoint on authorization server, using its own client ID and secret.

  3. Kafka client session is established if the token is valid.

Client using long-lived access token, with broker performing fast local validation

Client using long-lived access token with broker performing fast local validation

  1. Kafka client authenticates with the Kafka broker using the SASL OAUTHBEARER mechanism to pass the long-lived access token.

  2. Kafka broker validates the access token locally using JWT token signature check, and local token introspection.

Warning
Fast local JWT token signature validation is suitable only for short-lived tokens as there is no check with the authorization server if a token has been revoked. Token expiration is written into the token, but revocation can happen at any time, so cannot be accounted for without contacting the authorization server. Any issued token would be considered valid until it expires.

4.4.6. Configuring OAuth 2.0 authentication

OAuth 2.0 is used for interaction between Kafka clients and Strimzi components.

In order to use OAuth 2.0 for Strimzi, you must:

Configuring an OAuth 2.0 authorization server

This procedure describes in general what you need to do to configure an authorization server for integration with Strimzi.

These instructions are not product specific.

The steps are dependent on the chosen authorization server. Consult the product documentation for the authorization server for information on how to set up OAuth 2.0 access.

Note
If you already have an authorization server deployed, you can skip the deployment step and use your current deployment.
Procedure
  1. Deploy the authorization server to your cluster.

  2. Access the CLI or admin console for the authorization server to configure OAuth 2.0 for Strimzi.

    Now prepare the authorization server to work with Strimzi.

  3. Configure a kafka-broker client.

  4. Configure clients for each Kafka client component of your application.

What to do next

After deploying and configuring the authorization server, configure the Kafka brokers to use OAuth 2.0.

Configuring OAuth 2.0 support for Kafka brokers

This procedure describes how to configure Kafka brokers so that the broker listeners are enabled to use OAuth 2.0 authentication using an authorization server.

We advise use of OAuth 2.0 over an encrypted interface through configuration of TLS listeners. Plain listeners are not recommended.

If the authorization server is using certificates signed by the trusted CA and matching the OAuth 2.0 server hostname, TLS connection works using the default settings. Otherwise, you may need to configure the truststore with prober certificates or disable the certificate hostname validation.

When configuring the Kafka broker you have two options for the mechanism used to validate the access token during OAuth 2.0 authentication of the newly connected Kafka client:

Before you start

For more information on the configuration of OAuth 2.0 authentication for Kafka broker listeners, see:

Prerequisites
  • Strimzi and Kafka are running

  • An OAuth 2.0 authorization server is deployed

Procedure
  1. Update the Kafka broker configuration (Kafka.spec.kafka) of your Kafka resource in an editor.

    kubectl edit kafka my-cluster
  2. Configure the Kafka broker listeners configuration.

    The configuration for each type of listener does not have to be the same, as they are independent.

    The examples here show the configuration options as configured for external listeners.

    Example 1: Configuring fast local JWT token validation
    #...
    - name: external
      port: 9094
      type: loadbalancer
      tls: true
      authentication:
        type: oauth (1)
        validIssuerUri: <https://<auth-server-address>/auth/realms/external> (2)
        jwksEndpointUri: <https://<auth-server-address>/auth/realms/external/protocol/openid-connect/certs> (3)
        userNameClaim: preferred_username (4)
        maxSecondsWithoutReauthentication: 3600 (5)
        tlsTrustedCertificates: (6)
        - secretName: oauth-server-cert
          certificate: ca.crt
        disableTlsHostnameVerification: true (7)
        jwksExpirySeconds: 360 (8)
        jwksRefreshSeconds: 300 (9)
        jwksMinRefreshPauseSeconds: 1 (10)
    1. Listener type set to oauth.

    2. URI of the token issuer used for authentication.

    3. URI of the JWKS certificate endpoint used for local JWT validation.

    4. The token claim (or key) that contains the actual user name in the token. The user name is the principal used to identify the user. The userNameClaim value will depend on the authentication flow and the authorization server used.

    5. (Optional) Activates the Kafka re-authentication mechanism that enforces session expiry to the same length of time as the access token. If the specified value is less than the time left for the access token to expire, then the client will have to re-authenticate before the actual token expiry. By default, the session does not expire when the access token expires, and the client does not attempt re-authentication.

    6. (Optional) Trusted certificates for TLS connection to the authorization server.

    7. (Optional) Disable TLS hostname verification. Default is false.

    8. The duration the JWKS certificates are considered valid before they expire. Default is 360 seconds. If you specify a longer time, consider the risk of allowing access to revoked certificates.

    9. The period between refreshes of JWKS certificates. The interval must be at least 60 seconds shorter than the expiry interval. Default is 300 seconds.

    10. The minimum pause in seconds between consecutive attempts to refresh JWKS public keys. When an unknown signing key is encountered, the JWKS keys refresh is scheduled outside the regular periodic schedule with at least the specified pause since the last refresh attempt. The refreshing of keys follows the rule of exponential backoff, retrying on unsuccessful refreshes with ever increasing pause, until it reaches jwksRefreshSeconds. The default value is 1.

    Example 2: Configuring token validation using an introspection endpoint
    - name: external
      port: 9094
      type: loadbalancer
      tls: true
      authentication:
        type: oauth
        validIssuerUri: <https://<auth-server-address>/auth/realms/external>
        introspectionEndpointUri: <https://<auth-server-address>/auth/realms/external/protocol/openid-connect/token/introspect> (1)
        clientId: kafka-broker (2)
        clientSecret: (3)
          secretName: my-cluster-oauth
          key: clientSecret
        userNameClaim: preferred_username (4)
        maxSecondsWithoutReauthentication: 3600 (5)
    1. URI of the token introspection endpoint.

    2. Client ID to identify the client.

    3. Client Secret and client ID is used for authentication.

    4. The token claim (or key) that contains the actual user name in the token. The user name is the principal used to identify the user. The userNameClaim value will depend on the authorization server used.

    5. (Optional) Activates the Kafka re-authentication mechanism that enforces session expiry to the same length of time as the access token. If the specified value is less than the time left for the access token to expire, then the client will have to re-authenticate before the actual token expiry. By default, the session does not expire when the access token expires, and the client does not attempt re-authentication.

    Depending on how you apply OAuth 2.0 authentication, and the type of authorization server, there are additional (optional) configuration settings you can use:

      # ...
      authentication:
        type: oauth
        # ...
        checkIssuer: false (1)
        checkAudience: true (2)
        fallbackUserNameClaim: client_id (3)
        fallbackUserNamePrefix: client-account- (4)
        validTokenType: bearer (5)
        userInfoEndpointUri: https://OAUTH-SERVER-ADDRESS/auth/realms/external/protocol/openid-connect/userinfo (6)
        enableOauthBearer: false (7)
        enablePlain: true (8)
        tokenEndpointUri: https://OAUTH-SERVER-ADDRESS/auth/realms/external/protocol/openid-connect/token (9)
        customClaimCheck: "@.custom == 'custom-value'" (10)
        clientAudience: AUDIENCE (11)
        clientScope: SCOPE (12)
    1. If your authorization server does not provide an iss claim, it is not possible to perform an issuer check. In this situation, set checkIssuer to false and do not specify a validIssuerUri. Default is true.

    2. If your authorization server provides an aud (audience) claim, and you want to enforce an audience check, set checkAudience to true. Audience checks identify the intended recipients of tokens. As a result, the Kafka broker will reject tokens that do not have its clientId in their aud claim. Default is false.

    3. An authorization server may not provide a single attribute to identify both regular users and clients. When a client authenticates in its own name, the server might provide a client ID. When a user authenticates using a username and password, to obtain a refresh token or an access token, the server might provide a username attribute in addition to a client ID. Use this fallback option to specify the username claim (attribute) to use if a primary user ID attribute is not available.

    4. In situations where fallbackUserNameClaim is applicable, it may also be necessary to prevent name collisions between the values of the username claim, and those of the fallback username claim. Consider a situation where a client called producer exists, but also a regular user called producer exists. In order to differentiate between the two, you can use this property to add a prefix to the user ID of the client.

    5. (Only applicable when using introspectionEndpointUri) Depending on the authorization server you are using, the introspection endpoint may or may not return the token type attribute, or it may contain different values. You can specify a valid token type value that the response from the introspection endpoint has to contain.

    6. (Only applicable when using introspectionEndpointUri) The authorization server may be configured or implemented in such a way to not provide any identifiable information in an Introspection Endpoint response. In order to obtain the user ID, you can configure the URI of the userinfo endpoint as a fallback. The userNameClaim, fallbackUserNameClaim, and fallbackUserNamePrefix settings are applied to the response of userinfo endpoint.

    7. Set this to false`to disable the OAUTHBEARER mechanism on the listener. At least one of PLAIN or OAUTHBEARER has to be enabled. Default is `true.

    8. Set this to true to enable the PLAIN mechanism on the listener, which is supported by all clients on all platforms. The Kafka client has to enable the PLAIN mechanism and set the username and the password. This mechanism can be used to authenticate either by using the OAuth access token, or by using the OAuth client id and secret (client credentials). If the client sets password to start with the string $accessToken:, the password is interpreted as the access token on the server, and username as the account username, otherwise the user is interpreted as the client id, and password as the client secret. Default is false.

    9. This has to be set to support the client credentials authentication when enablePlain is set to true, as described in previous point.

    10. Additional custom rules can be imposed on the JWT access token during validation by setting this to a JsonPath filter query. If the access token does not contain the necessary data, it is rejected. When using the introspectionEndpointUri, the custom check is applied to the introspection endpoint response JSON.

    11. (Optional) An audience parameter passed to the token endpoint. An audience is used when obtaining an access token for inter-broker authentication. It is also used in the name of a client for OAuth 2.0 over PLAIN client authentication using a clientId and secret. This only affects the ability to obtain the token, and the content of the token, depending on the authorization server. It does not affect token validation rules by the listener.

    12. (Optional) A scope parameter passed to the token endpoint. A scope is used when obtaining an access token for inter-broker authentication. It is also used in the name of a client for OAuth 2.0 over PLAIN client authentication using a clientId and secret. This only affects the ability to obtain the token, and the content of the token, depending on the authorization server. It does not affect token validation rules by the listener.

  3. Save and exit the editor, then wait for rolling updates to complete.

  4. Check the update in the logs or by watching the pod state transitions:

    kubectl logs -f ${POD_NAME} -c ${CONTAINER_NAME}
    kubectl get pod -w

    The rolling update configures the brokers to use OAuth 2.0 authentication.

Configuring Kafka Java clients to use OAuth 2.0

This procedure describes how to configure Kafka producer and consumer APIs to use OAuth 2.0 for interaction with Kafka brokers.

Add a client callback plugin to your pom.xml file, and configure the system properties.

Prerequisites
  • Strimzi and Kafka are running

  • An OAuth 2.0 authorization server is deployed and configured for OAuth access to Kafka brokers

  • Kafka brokers are configured for OAuth 2.0

Procedure
  1. Add the client library with OAuth 2.0 support to the pom.xml file for the Kafka client:

    <dependency>
     <groupId>io.strimzi</groupId>
     <artifactId>kafka-oauth-client</artifactId>
     <version>{oauth-version}</version>
    </dependency>
  2. Configure the system properties for the callback:

    For example:

    System.setProperty(ClientConfig.OAUTH_TOKEN_ENDPOINT_URI, “https://<auth-server-address>/auth/realms/master/protocol/openid-connect/token”); (1)
    System.setProperty(ClientConfig.OAUTH_CLIENT_ID, "<client-name>"); (2)
    System.setProperty(ClientConfig.OAUTH_CLIENT_SECRET, "<client-secret>"); (3)
    1. URI of the authorization server token endpoint.

    2. Client ID, which is the name used when creating the client in the authorization server.

    3. Client secret created when creating the client in the authorization server.

  3. Enable the SASL OAUTHBEARER mechanism on a TLS encrypted connection in the Kafka client configuration:

    For example:

    props.put("sasl.jaas.config", "org.apache.kafka.common.security.oauthbearer.OAuthBearerLoginModule required;");
    props.put("security.protocol", "SASL_SSL"); (1)
    props.put("sasl.mechanism", "OAUTHBEARER");
    props.put("sasl.login.callback.handler.class", "io.strimzi.kafka.oauth.client.JaasClientOauthLoginCallbackHandler");
    1. Here we use SASL_SSL for use over TLS connections. Use SASL_PLAINTEXT over unencrypted connections.

  4. Verify that the Kafka client can access the Kafka brokers.

Configuring OAuth 2.0 for Kafka components

This procedure describes how to configure Kafka components to use OAuth 2.0 authentication using an authorization server.

You can configure authentication for:

  • Kafka Connect

  • Kafka MirrorMaker

  • Kafka Bridge

In this scenario, the Kafka component and the authorization server are running in the same cluster.

Before you start

For more information on the configuration of OAuth 2.0 authentication for Kafka components, see:

Prerequisites
  • Strimzi and Kafka are running

  • An OAuth 2.0 authorization server is deployed and configured for OAuth access to Kafka brokers

  • Kafka brokers are configured for OAuth 2.0

Procedure
  1. Create a client secret and mount it to the component as an environment variable.

    For example, here we are creating a client Secret for the Kafka Bridge:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Secret
    metadata:
     name: my-bridge-oauth
    type: Opaque
    data:
     clientSecret: MGQ1OTRmMzYtZTllZS00MDY2LWI5OGEtMTM5MzM2NjdlZjQw (1)
    1. The clientSecret key must be in base64 format.

  2. Create or edit the resource for the Kafka component so that OAuth 2.0 authentication is configured for the authentication property.

    For OAuth 2.0 authentication, you can use:

    • Client ID and secret

    • Client ID and refresh token

    • Access token

    • TLS

    For example, here OAuth 2.0 is assigned to the Kafka Bridge client using a client ID and secret, and TLS:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaBridge
    metadata:
      name: my-bridge
    spec:
      # ...
      authentication:
        type: oauth (1)
        tokenEndpointUri: https://<auth-server-address>/auth/realms/master/protocol/openid-connect/token (2)
        clientId: kafka-bridge
        clientSecret:
          secretName: my-bridge-oauth
          key: clientSecret
        tlsTrustedCertificates: (3)
        - secretName: oauth-server-cert
          certificate: tls.crt
    1. Authentication type set to oauth.

    2. URI of the token endpoint for authentication.

    3. Trusted certificates for TLS connection to the authorization server.

    Depending on how you apply OAuth 2.0 authentication, and the type of authorization server, there are additional configuration options you can use:

    # ...
    spec:
      # ...
      authentication:
        # ...
        disableTlsHostnameVerification: true (1)
        checkAccessTokenType: false (2)
        accessTokenIsJwt: false (3)
        scope: any (4)
        audience: kafka (5)
    1. (Optional) Disable TLS hostname verification. Default is false.

    2. If the authorization server does not return a typ (type) claim inside the JWT token, you can apply checkAccessTokenType: false to skip the token type check. Default is true.

    3. If you are using opaque tokens, you can apply accessTokenIsJwt: false so that access tokens are not treated as JWT tokens.

    4. (Optional) The scope for requesting the token from the token endpoint. An authorization server may require a client to specify the scope. In this case it is any.

    5. (Optional) The audience for requesting the token from the token endpoint. An authorization server may require a client to specify the audience. In this case it is kafka.

  3. Apply the changes to the deployment of your Kafka resource.

    kubectl apply -f your-file
  4. Check the update in the logs or by watching the pod state transitions:

    kubectl logs -f ${POD_NAME} -c ${CONTAINER_NAME}
    kubectl get pod -w

    The rolling updates configure the component for interaction with Kafka brokers using OAuth 2.0 authentication.

4.4.7. Authorization server examples

When choosing an authorization server, consider the features that best support configuration of your chosen authentication flow.

For the purposes of testing OAuth 2.0 with Strimzi, Keycloak and ORY Hydra were implemented as the OAuth 2.0 authorization server.

For more information, see:

4.5. Using OAuth 2.0 token-based authorization

If you are using OAuth 2.0 with Keycloak for token-based authentication, you can also use Keycloak to configure authorization rules to constrain client access to Kafka brokers. Authentication establishes the identity of a user. Authorization decides the level of access for that user.

Strimzi supports the use of OAuth 2.0 token-based authorization through Keycloak Keycloak Authorization Services, which allows you to manage security policies and permissions centrally.

Security policies and permissions defined in Keycloak are used to grant access to resources on Kafka brokers. Users and clients are matched against policies that permit access to perform specific actions on Kafka brokers.

Kafka allows all users full access to brokers by default, and also provides the AclAuthorizer plugin to configure authorization based on Access Control Lists (ACLs).

ZooKeeper stores ACL rules that grant or deny access to resources based on username. However, OAuth 2.0 token-based authorization with Keycloak offers far greater flexibility on how you wish to implement access control to Kafka brokers. In addition, you can configure your Kafka brokers to use OAuth 2.0 authorization and ACLs.

4.5.1. OAuth 2.0 authorization mechanism

OAuth 2.0 authorization in Strimzi uses Keycloak server Authorization Services REST endpoints to extend token-based authentication with Keycloak by applying defined security policies on a particular user, and providing a list of permissions granted on different resources for that user. Policies use roles and groups to match permissions to users. OAuth 2.0 authorization enforces permissions locally based on the received list of grants for the user from Keycloak Authorization Services.

Kafka broker custom authorizer

A Keycloak authorizer (KeycloakRBACAuthorizer) is provided with Strimzi. To be able to use the Keycloak REST endpoints for Authorization Services provided by Keycloak, you configure a custom authorizer on the Kafka broker.

The authorizer fetches a list of granted permissions from the authorization server as needed, and enforces authorization locally on the Kafka Broker, making rapid authorization decisions for each client request.

4.5.2. Configuring OAuth 2.0 authorization support

This procedure describes how to configure Kafka brokers to use OAuth 2.0 authorization using Keycloak Authorization Services.

Before you begin

Consider the access you require or want to limit for certain users. You can use a combination of Keycloak groups, roles, clients, and users to configure access in Keycloak.

Typically, groups are used to match users based on organizational departments or geographical locations. And roles are used to match users based on their function.

With Keycloak, you can store users and groups in LDAP, whereas clients and roles cannot be stored this way. Storage and access to user data may be a factor in how you choose to configure authorization policies.

Note
Super users always have unconstrained access to a Kafka broker regardless of the authorization implemented on the Kafka broker.
Prerequisites
  • Strimzi must be configured to use OAuth 2.0 with Keycloak for token-based authentication. You use the same Keycloak server endpoint when you set up authorization.

  • OAuth 2.0 authentication must be configured with the maxSecondsWithoutReauthentication option to enable re-authentication.

Procedure
  1. Access the Keycloak Admin Console or use the Keycloak Admin CLI to enable Authorization Services for the Kafka broker client you created when setting up OAuth 2.0 authentication.

  2. Use Authorization Services to define resources, authorization scopes, policies, and permissions for the client.

  3. Bind the permissions to users and clients by assigning them roles and groups.

  4. Configure the Kafka brokers to use Keycloak authorization by updating the Kafka broker configuration (Kafka.spec.kafka) of your Kafka resource in an editor.

    kubectl edit kafka my-cluster
  5. Configure the Kafka broker kafka configuration to use keycloak authorization, and to be able to access the authorization server and Authorization Services.

    For example:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    metadata:
      name: my-cluster
    spec:
      kafka:
        # ...
        authorization:
          type: keycloak (1)
          tokenEndpointUri: <https://<auth-server-address>/auth/realms/external/protocol/openid-connect/token> (2)
          clientId: kafka (3)
          delegateToKafkaAcls: false (4)
          disableTlsHostnameVerification: false (5)
          superUsers: (6)
          - CN=fred
          - sam
          - CN=edward
          tlsTrustedCertificates: (7)
          - secretName: oauth-server-cert
            certificate: ca.crt
          grantsRefreshPeriodSeconds: 60 (8)
          grantsRefreshPoolSize: 5 (9)
        #...
    1. Type keycloak enables Keycloak authorization.

    2. URI of the Keycloak token endpoint. For production, always use HTTPs.

    3. The client ID of the OAuth 2.0 client definition in Keycloak that has Authorization Services enabled. Typically, kafka is used as the ID.

    4. (Optional) Delegate authorization to Kafka AclAuthorizer if access is denied by Keycloak Authorization Services policies. Default is false.

    5. (Optional) Disable TLS hostname verification. Default is false.

    6. (Optional) Designated super users.

    7. (Optional) Trusted certificates for TLS connection to the authorization server.

    8. (Optional) The time between two consecutive grants refresh runs. That is the maximum time for active sessions to detect any permissions changes for the user on Keycloak. The default value is 60.

    9. (Optional) The number of threads to use to refresh (in parallel) the grants for the active sessions. The default value is 5.

  6. Save and exit the editor, then wait for rolling updates to complete.

  7. Check the update in the logs or by watching the pod state transitions:

    kubectl logs -f ${POD_NAME} -c kafka
    kubectl get pod -w

    The rolling update configures the brokers to use OAuth 2.0 authorization.

  8. Verify the configured permissions by accessing Kafka brokers as clients or users with specific roles, making sure they have the necessary access, or do not have the access they are not supposed to have.

4.5.3. Managing policies and permissions in Keycloak Authorization Services

This section describes the mappings between the Kafka authorization model and Keycloak Authorization Services model. The mappings are used in granting permissions to access Kafka.

Kafka authorization model for resources

The Kafka authorization model defines resource types, and the permissions available for each type. When an action is performed by a Kafka client on a broker, the broker uses a configured authorizer to check permissions, depending on the action performed and the resource type.

Kafka has five resource types for controlling access: Topic, Group, Cluster, TransactionalId, DelegationToken.

Each resource type has different permissions:

Topic:

  • Create

  • Write

  • Read

  • Delete

  • Describe

  • DescribeConfigs

  • Alter

  • AlterConfigs

Group:

  • Read

  • Describe

  • Delete

Cluster:

  • Create

  • Describe

  • Alter

  • DescribeConfigs

  • AlterConfigs

  • IdempotentWrite

  • ClusterAction

TransactionalId:

  • Describe

  • Write

DelegationToken:

  • Describe

Keycloak Authorization Services model for managing permissions

Keycloak Authorization Services use four concepts to define and grant permissions: resources, authorization scopes, policies, and permissions.

Resources

Resources are a set of resource definitions that are used to match permitted actions. For example, a resource can be an individual topic, or it can be a set of all topics with names that start with the same prefix. The resource definition has a set of available authorization scopes associated with it, which represent a set of all actions available on the particular resource. Often, only a subset of these actions is actually permitted.

Authorization scopes

Authorization scopes is simply a set of all available actions on all the different resource types. When defining a new resource, scopes are added from the set of all scopes.

Policies

Policies are rules that use criteria to match a list of accounts. Policies can match service accounts based on client id or roles, or user accounts based on username, groups, or roles.

Permissions

Permissions grant a subset of authorization scopes on a specific resource definition to a set of users.

Mapping Keycloak Authorization Services to the Kafka authorization model

Use Keycloak Authorization Services rules on the OAuth client that represents the Kafka Broker to grant Kafka permissions to users or service accounts. Typically, the OAuth client has kafka as its client id.

The OAuth 2.0 client definition must have the Authorization Enabled option activated.

All permissions exist within the scope of this OAuth 2.0 client, which means that if you have different Kafka clusters configured with different OAuth 2.0 client IDs they would each have a separate set of permissions even though they are part of the same realm.

When the Kafka client use the SASL OAUTHBEARER mechanism, the Keycloak authorizer (KeycloakRBACAuthorizer) retrieves the list of grants for the current session from the Keycloak server using the access token of the current session. This list of grants is the result of evaluating the Keycloak Authorization Services policies and permissions.

Introducing authorization scopes

Typically, an initial configuration involves uploading the authorization scopes to create a list of all the possible actions that can be performed on all the Kafka resource types. This step is performed only once, before defining any permissions. Alternatively, you can add authorization scopes manually.

The authorization scopes should contain all the possible Kafka permissions regardless of the resource type:

  • Create

  • Write

  • Read

  • Delete

  • Describe

  • Alter

  • DescribeConfig

  • AlterConfig

  • ClusterAction

  • IdempotentWrite

Defining resource patterns for permission checks

The resources use pattern names for pattern matching against the targeted resources when performing permission checks.

The general pattern is RESOURCE-TYPE:PATTERN-NAME.

The resource types mirror the Kafka authorization model. The pattern allows for the two matching options: exact matching (when the pattern does not end with *), and prefix matching (when the pattern ends with *).

Example patterns for resources
Topic:my-topic
Topic:orders-*
Group:orders-*
Cluster:*

In addition, the general pattern can be prefixed by kafka-cluster:CLUSTER-NAME followed by a comma, where the cluster name refers to the metadata.name in the Kafka custom resource.

Example patterns for resources with cluster prefix
kafka-cluster:my-cluster,Topic:*
kafka-cluster:*,Group:b_*

When the kafka-cluster prefix is not present it is assumed to be kafka-cluster:*.

When defining a resource, you can associate a list of possible authorization scopes relevant to the resource. Set whatever actions make sense for the targeted resource type.

While you may add any authorization scope to any resource, only the scopes supported by the resource type are considered for access control.

Policies

Policies are used to target permissions to one or more accounts. Targeting can refer to:

  • Specific user or service accounts

  • Realm roles or client roles

  • User groups

  • JavaScript rule to match a client IP address

A policy is given a unique name, and can be reused to target multiple permissions to multiple resources.

Defining permissions based on scopes, resources and policies

Use fine-grained permissions to pull together the policies, resources, and authorization scopes that grant access to users.

The name of each permission should clearly define what permissions it grants to which users.

For more information on how to configure permissions through Keycloak Authorization Services, see the authorization example.

Example permissions required for operations on Kafka

The following examples demonstrate the user permissions required for performing common operations on Kafka.

Creating a topic

To create a topic, the Create permission is required for the specific topic, or for Cluster:kafka-cluster.

bin/kafka-topics.sh --create --topic my-topic \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Listing the topic

If a user has Describe permission on the topic, the topic is listed.

bin/kafka-topics.sh --list \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Displaying the topic details

To display topic details, Describe and DescribeConfigs permissions are required on the topic.

bin/kafka-topics.sh --describe --topic my-topic \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Producing to the topic

To produce to the topic, Describe and Write permissions are required on the topic. If topic has not yet been created, and autocreation is enabled, the permissions to create the topic are required.

bin/kafka-console-producer.sh  --topic my-topic \
  --broker-list my-cluster-kafka-bootstrap:9092 --producer.config=/tmp/config.properties
Consuming from the topic

To consume from the topic, Describe and Read permissions are required on the topic. Consuming from the topic normally relies on storing the consumer offsets in a consumer group, which requires additional Describe and Read permissions on the consumer group.

Two resources are needed for matching. For example:

Topic:my-topic
Group:my-group-*
bin/kafka-console-consumer.sh --topic my-topic --group my-group-1 --from-beginning \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --consumer.config /tmp/config.properties
Producing to the topic using an idempotent producer

Besides needing the permissions for standard producing to the topic, an additional IdempotentWrite permission is required on the Cluster resource.

Two resources are needed for matching. For example:

Topic:my-topic
Cluster:kafka-cluster
Listing consumer groups

When listing consumer groups, only the groups on which the user has Describe permissions are returned. Alternatively, if the user has Describe permission on the Cluster:kafka-cluster, all the consumer groups are returned.

bin/kafka-consumer-groups.sh --list \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Displaying the consumer group details

To display the consumer group details, Describe permission is required on the group, and on the topic associated with the group.

bin/kafka-consumer-groups.sh --describe --group my-group-1 \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Changing the topic configuration

To change the topic configuration, Describe and Alter permissions are required on the topic.

bin/kafka-topics.sh --alter --topic my-topic --partitions 2 \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Displaying the Kafka broker configuration

To be able to use kafka-configs.sh to get the broker configuration, DescribeConfigs permission is required on the Cluster:kafka-cluster.

bin/kafka-configs.sh --entity-type brokers --entity-name 0 --describe --all \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Changing the Kafka broker configuration

To change the Kafka broker configuration, DescribeConfigs and AlterConfigs permissions are required on Cluster:kafka-cluster.

bin/kafka-configs --entity-type brokers --entity-name 0 --alter --add-config log.cleaner.threads=2 \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Deleting a topic

To delete the topic, Describe and Delete permissions are required on the topic.

bin/kafka-topics.sh --delete --topic my-topic \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config=/tmp/config.properties
Selecting a leader partition

To run leader selection for topic partitions, Alter permission is required on the Cluster:kafka-cluster.

bin/kafka-leader-election.sh --topic my-topic --partition 0 --election-type PREFERRED  /
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --admin.config /tmp/config.properties
Reassigning partitions

To generate a partition reassignment file, Describe permissions are required on the topics involved.

bin/kafka-reassign-partitions.sh --topics-to-move-json-file /tmp/topics-to-move.json --broker-list "0,1" --generate \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config /tmp/config.properties > /tmp/partition-reassignment.json

To execute the partition reassignment, Describe and Alter permissions are required on Cluster:kafka-cluster, and Describe permissions are required on the topics involved.

bin/kafka-reassign-partitions.sh --reassignment-json-file /tmp/partition-reassignment.json --execute \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config /tmp/config.properties

To verify partition reassignment, Describe, and AlterConfigs permissions are required on Cluster:kafka-cluster, and on each of the topics involved.

bin/kafka-reassign-partitions.sh --reassignment-json-file /tmp/partition-reassignment.json --verify \
  --bootstrap-server my-cluster-kafka-bootstrap:9092 --command-config /tmp/config.properties

4.5.4. Example authorization rules configuration using Keycloak Authorization Services

This section provides an end-to-end example of using Keycloak Authorization Services to configure authorization rules for use with keycloak authorization.

First, we use the Admin Console to import a preconfigured realm into a running Keycloak server.

Second, we configure a Kafka cluster to use the preconfigured realm and then deploy the cluster to Kubernetes.

Third, in the Admin Console, we demonstrate the parts of the Keycloak user interface that configure authorization rules.

Finally, we use various Kafka CLIs to show how access to Kafka resources is limited based on the permissions granted to different personal and service accounts.

Security considerations

This example explains how to use Keycloak Authorization Services with keycloak authorization. Although TLS connectivity is used where necessary, the server certificates are auto-generated, and TLS termination for the Keycloak server is on a reverse proxy server using an Ingress, not on the Keycloak server pod. Take care when creating and reviewing any production configuration. Keycloak requires Kafka clients to use the same hostname and port as the Kafka brokers. This example assumes that the Ingress hostname is visible to the pods inside the Kubernetes cluster. In your cluster, the Ingress might be visible to external Kafka clients, but not to the pods. Alternatively, you can deploy the Keycloak server to a different Kubernetes cluster than the Cluster Operator.

Token-based authorization with Keycloak Authorization Services

When using Keycloak with oauth authentication, the access token is available, so it is possible to use centrally managed authorization rules to enforce access restrictions onto Kafka Clients. Strimzi provides support for Keycloak Authorization Services in the form of keycloak authorization, which allows you to use Keycloak to manage fine grained permissions to Kafka resources.

When using keycloak authorization, a custom authorizer is configured on the Kafka broker that uses Authorization Services REST endpoints available on Keycloak, which provide a list of granted permissions on resources for authenticated users. The list of grants (permissions) is fetched as the first action after an authenticated session is established by the Kafka client, and then regularly refreshed in the background. Grants are cached and enforced locally on the Kafka broker for each user session to provide fast authorization decisions. Because they are periodically refreshed, any changes to the grants on the Keycloak server are detected and enforced.

Provisioning the Keycloak server

Install the Keycloak server using the Keycloak Operator as described in Installing the Keycloak Operator in the Keycloak documentation.

Uploading the demo realm

When a Keycloak instance is running, upload the example realm to Keycloak using the Admin Console.

First, determine the external hostname to access the Admin Console.

For example, if the Keycloak Operator installed the Keycloak server in the sso namespace, run the following command:

NS=sso
kubectl get ingress keycloak -n $NS

Second, get the password for the admin user.

If you used Keycloak Operator, the password is stored as a secret, so you first have to identify the name of the secret. You can inspect the yaml for the Keycloak instance. For example:

kubectl get -n $NS pod keycloak-0 -o yaml | less

Search for KEYCLOAK_PASSWORD and note the secretKeyRef.name (usually credential-keycloak or similar).

You can then obtain the clear text password by running:

SECRET_NAME=credential-keycloak
kubectl get -n $NS secret $SECRET_NAME -o yaml | grep PASSWORD | awk '{print $2}' | base64 -D

Log in to the Admin Console with the username admin and the password you obtained earlier. You must use https://HOST to access the Kubernetes ingress.

When you are logged in, click Add Realm to import the example realm.

Select the examples/security/keycloak-authorization/kafka-authz-realm.json file from your disk, and then click Create.

You should now have kafka-authz as your current realm in the Admin Console.

Deploying the minimal Kafka cluster

We assume that the Strimzi Cluster Operator has already been installed on the Kubernetes cluster.

Our Kafka cluster has to be properly configured with authentication and authorization settings to connect to our Keycloak instance.

Assuming the Keycloak Operator was used to install the Keycloak server instance, and that sso namespace was used, we can determine the external hostname:

NS=sso
kubectl get ingress keycloak -n $NS

Kubernetes ingress uses a secure connection (HTTPS), which means that we need to prepare a certificate truststore to make it possible for Kafka brokers to communicate with the Keycloak server.

SSO_HOST=<sso_hostname>
SSO_HOST_PORT=$SSO_HOST:443
STOREPASS=storepass

echo "Q" | openssl s_client -showcerts -connect $SSO_HOST_PORT 2>/dev/null | awk ' /BEGIN CERTIFICATE/,/END CERTIFICATE/ { print $0 } ' > /tmp/sso.crt

Now we need to deploy the certificate to Kubernetes as a secret:

kubectl create secret generic oauth-server-cert --from-file=/tmp/sso.crt -n $NS

Set the hostname as an ENV var before deploying the prepared Kafka custom resource for the example:

SSO_HOST=<sso_hostname>
EXAMPLES_DIR=examples
cat $EXAMPLES_DIR/security/keycloak-authorization/kafka-ephemeral-oauth-single-keycloak-authz.yaml | sed -E 's#\${SSO_HOST}'"#$SSO_HOST#" | kubectl create -n $NS -f -
Using the Keycloak Admin Console to configure authorization

After logging into the Admin Console, the default view displays the Master realm. For this example, locate and select the example kafka-authz realm that you uploaded earlier.

Initially, the Realm Settings section is selected, but you can navigate to Groups, Roles, Clients and Users.

Under Groups, you can view user groups and set user permissions. Groups are sets of users with a name assigned. Typically, they are used to compartmentalize users into geographical, organizational or departmental units.

In Keycloak, groups can be stored in an LDAP identity provider. You can make a user a member of a group through a custom LDAP server admin user interface, for example, to grant permissions on Kafka resources.

Under Users, you can view all defined users. For this example, alice and bob are defined. alice is a member of the ClusterManager Group, and bob is a member of ClusterManager-my-cluster Group. In Keycloak, users can be stored in an LDAP identity provider.

Under Roles, you can view the realm roles to mark users or clients as having some permissions. Roles are a concept analogous to groups. They are usually used to tag users with organizational roles and have the requisite permissions. Roles cannot be stored in an LDAP identity provider. If LDAP is a requirement, you can use groups instead, and add Keycloak roles to the groups so that when users are assigned a group they also get a corresponding role.

Under Clients, you can view the additional client configurations. For this example, kafka, kafka-cli, team-a-client, and team-b-client are configured. The client with client id kafka is used by Kafka brokers to perform the necessary OAuth 2.0 communication for access token validation. This client also contains the Authorization Services resource definitions, policies and authorization scopes used to perform authorization on the Kafka brokers.

The client with client id kafka-cli is a public client that can be used by the Kafka command line tools when authenticating with username and password to obtain an access token or a refresh token.

Clients team-a-client, and team-b-client are confidential clients representing services with partial access to certain Kafka topics.

The authorization configuration is defined in the kafka client from the Authorization tab, which becomes visible when Authorization Enabled is switched on from the Settings tab.

Defining Authorization Services for access control

Keycloak Authorization Services use authorization scopes, policies and permissions to define and apply access control to resources, as explained in Keycloak Authorization Services model for managing permissions.

From Authorization / Permissions you can see the granted permissions that use resources and policies defined from other Resources and Policies tabs. For example, the kafka client has the following permissions:

Dev Team A can write to topics that start with x_ on any cluster
Dev Team B can read from topics that start with x_ on any cluster
Dev Team B can update consumer group offsets that start with x_ on any cluster
ClusterManager of my-cluster Group has full access to cluster config on my-cluster
ClusterManager of my-cluster Group has full access to consumer groups on my-cluster
ClusterManager of my-cluster Group has full access to topics on my-cluster

Dev Team A can write to topics that start with x_ on any cluster combines a resource called Topic:x_*, scopes Describe and Write, and Dev Team A policy. The Dev Team A policy matches all users that have a realm role called Dev Team A.

Dev Team B can read from topics that start with x_ on any cluster combines Topic:x_*, and Group:x_* resources, scopes Describe and Read, and Dev Team B policy. The Dev Team B policy matches all users that have a realm role called Dev Team B. Matching users and clients have the ability to read from topics, and update the consumed offsets for topics and consumer groups that have names starting with x_.

Targeting permissions using group or role policies

In Keycloak, confidential clients with service accounts enabled can authenticate to the server in their own name using a client ID and a secret. This is convenient for microservices which typically act in their own name, and not as agents of a particular user (like a web site would, for example). Service accounts can have roles assigned like regular users. They cannot, however, have groups assigned. As a consequence, if you want to target permissions to microservices using service accounts, you cannot use group policies, and should instead use role policies. Conversely, if you want to limit certain permissions only to regular user accounts where authentication with username and password is required, you can achieve that as a side effect of using the group policies, rather than the role policies. That is what is used for permissions that start with ClusterManager. Performing cluster management is usually done interactively - in person - using CLI tools. It makes sense to require the user to log in before using the resulting access token to authenticate to the Kafka broker. In this case, the access token represents the specific user, rather than the client application.

Authorization in action using CLI clients

Before checking authorized access to Kafka, the imported authorization rules must be present in the Admin Console.

From Clients  kafka  Authorization  Settings Decision Strategy must be set to Affirmative, and NOT to Unanimous. The Resources, Authorization claims, Policies, and Permissions tabs must contain the authorization content.

With the configuration in place, we can check access to Kafka by using a producer and consumer to create topics using different user and service accounts.

First, a new interactive pod container is run using a Strimzi Kafka image to connect to a running Kafka broker.

NS=sso
kubectl run -ti --restart=Never --image=quay.io/strimzi/kafka:0.24.0-kafka-2.8.0 kafka-cli -n $NS -- /bin/sh
Note
If kubectl times out waiting on the image download, subsequent attempts may result in an AlreadyExists error.

You can attach to the existing pod by running:

kubectl attach -ti kafka-cli -n $NS

To produce messages as client team-a-client, we prepare a Kafka client configuration file. We use the SASL/OAUTHBEARER mechanism with a Client ID and Client Secret, which means the client first connects to the Keycloak server to obtain an access token. The client then connects to the Kafka broker and uses the obtained access token to authenticate.

We need to prepare and configure the truststore for TLS connections to work.

First, we use the external hostname exposing the Keycloak to obtain the certificate.

SSO_HOST=<sso_hostname>
SSO_HOST_PORT=$SSO_HOST:443
STOREPASS=storepass

echo "Q" | openssl s_client -showcerts -connect $SSO_HOST_PORT 2>/dev/null | awk ' /BEGIN CERTIFICATE/,/END CERTIFICATE/ { print $0 } ' > /tmp/sso.crt

keytool -keystore /tmp/truststore.p12 -storetype pkcs12 -alias sso -storepass $STOREPASS -import -file /tmp/sso.crt -noprompt

Then, we add to the same truststore the certificate for the Kafka broker, which we obtain using the my-cluster-kafka-bootstrap as a hostname and tls listener port (9093):

KAFKA_HOST_PORT=my-cluster-kafka-bootstrap:9093
STOREPASS=storepass

echo "Q" | openssl s_client -showcerts -connect $KAFKA_HOST_PORT 2>/dev/null | awk ' /BEGIN CERTIFICATE/,/END CERTIFICATE/ { print $0 } ' > /tmp/my-cluster-kafka.crt

keytool -keystore /tmp/truststore.p12 -storetype pkcs12 -alias my-cluster-kafka -storepass $STOREPASS -import -file /tmp/my-cluster-kafka.crt -noprompt

Finally, we prepare the Kafka Client configuration parameters:

SSO_HOST=<sso_hostname>

cat > /tmp/team-a-client.properties << EOF
security.protocol=SASL_SSL
ssl.truststore.location=/tmp/truststore.p12
ssl.truststore.password=$STOREPASS
ssl.truststore.type=PKCS12
sasl.mechanism=OAUTHBEARER
sasl.jaas.config=org.apache.kafka.common.security.oauthbearer.OAuthBearerLoginModule required \
  oauth.client.id="team-a-client" \
  oauth.client.secret="team-a-client-secret" \
  oauth.ssl.truststore.location="/tmp/truststore.p12" \
  oauth.ssl.truststore.password="$STOREPASS" \
  oauth.ssl.truststore.type="PKCS12" \
  oauth.token.endpoint.uri="https://$SSO_HOST/auth/realms/kafka-authz/protocol/openid-connect/token" ;
sasl.login.callback.handler.class=io.strimzi.kafka.oauth.client.JaasClientOauthLoginCallbackHandler
EOF

The roles assigned to a client, such as the Dev Team A realm role assigned to the team-a-client service account, are presented on the Service Account Roles tab for Clients in the Admin Console.

We can use this configuration from the Kafka CLI to produce and consume messages, and perform other administration tasks.

Producing messages with authorized access

The team-a-client configuration is used to produce messages to topics that start with a_ or x_. The first command will result in an error as it tries to write to topic my-topic:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic my-topic \
  --producer.config=/tmp/team-a-client.properties
First message

A Not authorized to access topics: [my-topic] error is returned when trying to push the first message.

team-a-client has a Dev Team A role that gives it permission to perform any supported actions on topics that start with a_, but can only write to topics that start with x_. The topic named my-topic matches neither of those rules.

The team-a-client configuration is then used to produce messages to topic a_messages:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic a_messages \
  --producer.config /tmp/team-a-client.properties
First message
Second message

The messages are pushed out successfully, and in the Kafka container log there is DEBUG level output saying Authorization GRANTED.

Use CTRL-C to exit the CLI application.

You can see the Kafka container log by running:

kubectl logs my-cluster-kafka-0 -f -n $NS
Consuming messages with authorized access

The team-a-client configuration can be used to consume messages from topic a_messages. However, the following command will produce an error:

bin/kafka-console-consumer.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --topic a_messages \
  --from-beginning --consumer.config /tmp/team-a-client.properties

An error is returned as the Dev Team A role for team-a-client only has access to consumer groups that have names starting with a_. The team-a-client properties are updated to specify the custom consumer group it is permitted to use:

bin/kafka-console-consumer.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --topic a_messages \
  --from-beginning --consumer.config /tmp/team-a-client.properties --group a_consumer_group_1

This time the consumer receives all the messages from the a_messages topic.

Administering Kafka with authorized access

The team-a-client is an account without any cluster-level access, but it can be used with some administrative operations.

Listing topics returns the a_messages topic:

bin/kafka-topics.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/team-a-client.properties --list

Listing consumer groups returns the a_consumer_group_1 consumer group:

bin/kafka-consumer-groups.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/team-a-client.properties --list

Fetching the default cluster configuration fails cluster authorization, because the operation requires cluster level permissions that team-a-client does not have:

bin/kafka-configs.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/team-a-client.properties \
  --entity-type brokers --describe --entity-default
Using clients with different permissions

As with team-a-client, we prepare a Kafka client configuration file with authentication parameters for team-b-client:

cat > /tmp/team-b-client.properties << EOF
security.protocol=SASL_SSL
ssl.truststore.location=/tmp/truststore.p12
ssl.truststore.password=$STOREPASS
ssl.truststore.type=PKCS12
sasl.mechanism=OAUTHBEARER
sasl.jaas.config=org.apache.kafka.common.security.oauthbearer.OAuthBearerLoginModule required \
  oauth.client.id="team-b-client" \
  oauth.client.secret="team-b-client-secret" \
  oauth.ssl.truststore.location="/tmp/truststore.p12" \
  oauth.ssl.truststore.password="$STOREPASS" \
  oauth.ssl.truststore.type="PKCS12" \
  oauth.token.endpoint.uri="https://$SSO_HOST/auth/realms/kafka-authz/protocol/openid-connect/token" ;
sasl.login.callback.handler.class=io.strimzi.kafka.oauth.client.JaasClientOauthLoginCallbackHandler
EOF

The team-b-client Keycloak configuration includes a Dev Team B realm role and permissions that start with Dev Team B …​. These match the users and service accounts that have the Dev Team B realm role assigned to them. The Dev Team B users have full access to topics beginning with b_ on the Kafka cluster my-cluster, and read access on topics that start with x_.

The team-b-client configuration is used to produce messages to topics that start with b_. Writing to topic a_messages will result in error:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic a_messages \
  --producer.config /tmp/team-b-client.properties
Message 1

A Not authorized to access topics: [a_messages] error is returned when trying to push the first message, as expected, so we switch to topic b_messages:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic b_messages \
  --producer.config /tmp/team-b-client.properties
Message 1
Message 2
Message 3

Producing messages to topic b_messages is authorized and successful.

We switch again, but this time to a topic that team-b-client can only read from, topic x_messages:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --producer.config /tmp/team-b-client.properties
Message 1

A Not authorized to access topics: [x_messages] error is returned, as expected, so we switch to team-a-client:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --producer.config /tmp/team-a-client.properties
Message 1

A Not authorized to access topics: [x_messages] error is returned again. Though team-a-client can write to the x_messages topic, it does not have a permission to create a topic if it does not yet exist.

Before team-a-client can write to the x_messages topic, an admin power user must create it with the correct configuration, such as the number of partitions and replicas.

Managing Kafka with an authorized admin

Admin user alice is created in Keycloak with full access to manage everything on any Kafka cluster.

We can authenticate as alice by using curl and performing a password grant authentication to obtain a refresh token, which we can then use to configure the Kafka client.

USERNAME=alice
PASSWORD=alice-password

GRANT_RESPONSE=$(curl -X POST "https://$SSO_HOST/auth/realms/kafka-authz/protocol/openid-connect/token" -H 'Content-Type: application/x-www-form-urlencoded' -d "grant_type=password&username=$USERNAME&password=$PASSWORD&client_id=kafka-cli&scope=offline_access" -s -k)

REFRESH_TOKEN=$(echo $GRANT_RESPONSE | awk -F "refresh_token\":\"" '{printf $2}' | awk -F "\"" '{printf $1}')

The refresh token in this case is an offline token which is a long-lived refresh token that does not expire.

The configuration file for alice:

cat > /tmp/alice.properties << EOF
security.protocol=SASL_SSL
ssl.truststore.location=/tmp/truststore.p12
ssl.truststore.password=$STOREPASS
ssl.truststore.type=PKCS12
sasl.mechanism=OAUTHBEARER
sasl.jaas.config=org.apache.kafka.common.security.oauthbearer.OAuthBearerLoginModule required \
  oauth.refresh.token="$REFRESH_TOKEN" \
  oauth.client.id="kafka-cli" \
  oauth.ssl.truststore.location="/tmp/truststore.p12" \
  oauth.ssl.truststore.password="$STOREPASS" \
  oauth.ssl.truststore.type="PKCS12" \
  oauth.token.endpoint.uri="https://$SSO_HOST/auth/realms/kafka-authz/protocol/openid-connect/token" ;
sasl.login.callback.handler.class=io.strimzi.kafka.oauth.client.JaasClientOauthLoginCallbackHandler
EOF

The kafka-cli public client is used for the oauth.client.id in the sasl.jaas.config. Since that is a public client it does not require a Secret. We can use this because we authenticate with a token directly. In this case, the refresh token requests an access token behind the scenes, which is then sent to the Kafka broker for authentication. The refresh token has already been authenticated.

User alice has permission to create the x_messages topic:

bin/kafka-topics.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/alice.properties \
  --topic x_messages --create --replication-factor 1 --partitions 1

User alice can list all the topics, whereas team-a-client and team-b-client can only list the topics they have access to:

bin/kafka-topics.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/alice.properties --list
bin/kafka-topics.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/team-a-client.properties --list
bin/kafka-topics.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/team-b-client.properties --list

The Dev Team A and Dev Team B roles both have Describe permission on topics that start with x_, but they cannot see the other team’s topics because they do not have Describe permissions on them.

The team-a-client can now successfully produce to the x_messages topic:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --producer.config /tmp/team-a-client.properties
Message 1
Message 2
Message 3

As expected, team-b-client still cannot produce to the x_messages topic, and the following operation returns an error:

bin/kafka-console-producer.sh --broker-list my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --producer.config /tmp/team-b-client.properties
Message 4
Message 5

However, due to its Keycloak settings team-b-client can consume messages from the x_messages topic:

bin/kafka-console-consumer.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --from-beginning --consumer.config /tmp/team-b-client.properties --group x_consumer_group_b

Conversely, even though team-a-client can write to topic x_messages, the following read request returns a Not authorized to access group: x_consumer_group_a error:

bin/kafka-console-consumer.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --from-beginning --consumer.config /tmp/team-a-client.properties --group x_consumer_group_a

A consumer group that begins with a_ is used in the next read request:

bin/kafka-console-consumer.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --from-beginning --consumer.config /tmp/team-a-client.properties --group a_consumer_group_a

An error is still returned, but this time it is Not authorized to access topics: [x_messages].

Dev Team A has no Read access on topics that start with 'x_'.

User alice can read from or write to any topic:

bin/kafka-console-consumer.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --topic x_messages \
  --from-beginning --consumer.config /tmp/alice.properties

User alice can also read the cluster configuration (which in this case is empty):

bin/kafka-configs.sh --bootstrap-server my-cluster-kafka-bootstrap:9093 --command-config /tmp/alice.properties \
  --entity-type brokers --describe --entity-default

5. Using Strimzi Operators

Use the Strimzi operators to manage your Kafka cluster, and Kafka topics and users.

5.1. Using the Cluster Operator

The Cluster Operator is used to deploy a Kafka cluster and other Kafka components.

The Cluster Operator is deployed using YAML installation files.

Note
On OpenShift, a Kafka Connect deployment can incorporate a Source2Image feature to provide a convenient way to add additional connectors.
Additional resources

5.1.1. Cluster Operator configuration

You can configure the Cluster Operator using supported environment variables, and through its logging configuration.

The environment variables relate to container configuration for the deployment of the Cluster Operator image. For more information on image configuration, see, image.

STRIMZI_NAMESPACE

A comma-separated list of namespaces that the operator should operate in. When not set, set to empty string, or set to *, the Cluster Operator will operate in all namespaces. The Cluster Operator deployment might use the Kubernetes Downward API to set this automatically to the namespace the Cluster Operator is deployed in.

Example configuration for Cluster Operator namespaces
env:
  - name: STRIMZI_NAMESPACE
    valueFrom:
      fieldRef:
        fieldPath: metadata.namespace
STRIMZI_FULL_RECONCILIATION_INTERVAL_MS

Optional, default is 120000 ms. The interval between periodic reconciliations, in milliseconds.

STRIMZI_OPERATION_TIMEOUT_MS

Optional, default 300000 ms. The timeout for internal operations, in milliseconds. This value should be increased when using Strimzi on clusters where regular Kubernetes operations take longer than usual (because of slow downloading of Docker images, for example).

STRIMZI_OPERATIONS_THREAD_POOL_SIZE

Optional, default 10 The worker thread pool size, which is used for various asynchronous and blocking operations that are run by the cluster operator.

STRIMZI_OPERATOR_NAMESPACE

The name of the namespace where the Strimzi Cluster Operator is running. Do not configure this variable manually. Use the Kubernetes Downward API.

env:
  - name: STRIMZI_OPERATOR_NAMESPACE
    valueFrom:
      fieldRef:
        fieldPath: metadata.namespace
STRIMZI_OPERATOR_NAMESPACE_LABELS

Optional. The labels of the namespace where the Strimzi Cluster Operator is running. Namespace labels are used to configure the namespace selector in network policies to allow the Strimzi Cluster Operator to only have access to the operands from the namespace with these labels. When not set, the namespace selector in network policies is configured to allow access to the Strimzi Cluster Operator from any namespace in the Kubernetes cluster.

env:
  - name: STRIMZI_OPERATOR_NAMESPACE_LABELS
    value: label1=value1,label2=value2
STRIMZI_CUSTOM_RESOURCE_SELECTOR

Optional. Specifies label selector used to filter the custom resources handled by the operator. The operator will operate only on those custom resources which will have the specified labels set. Resources without these labels will not be seen by the operator. The label selector applies to Kafka, KafkaConnect, KafkaConnectS2I, KafkaBridge, KafkaMirrorMaker, and KafkaMirrorMaker2 resources. KafkaRebalance and KafkaConnector resources will be operated only when their corresponding Kafka and Kafka Connect clusters have the matching labels.

env:
  - name: STRIMZI_CUSTOM_RESOURCE_SELECTOR
    value: label1=value1,label2=value2
STRIMZI_LABELS_EXCLUSION_PATTERN

Optional, default regex pattern is ^app.kubernetes.io/(?!part-of).*. Specifies regex exclusion pattern used to filter labels propagation from the main custom resource to its subresources. The labels exclusion filter is not applied to labels in template sections such as spec.kafka.template.pod.metadata.labels.

env:
  - name: STRIMZI_LABELS_EXCLUSION_PATTERN
    value: "^key1.*"
STRIMZI_KAFKA_IMAGES

Required. This provides a mapping from Kafka version to the corresponding Docker image containing a Kafka broker of that version. The required syntax is whitespace or comma separated <version>=<image> pairs. For example 2.7.0=quay.io/strimzi/kafka:0.24.0-kafka-2.7.0, 2.8.0=quay.io/strimzi/kafka:0.24.0-kafka-2.8.0. This is used when a Kafka.spec.kafka.version property is specified but not the Kafka.spec.kafka.image in the Kafka resource.

STRIMZI_DEFAULT_KAFKA_INIT_IMAGE

Optional, default quay.io/strimzi/operator:0.24.0. The image name to use as default for the init container started before the broker for initial configuration work (that is, rack support), if no image is specified as the kafka-init-image in the Kafka resource.

STRIMZI_KAFKA_CONNECT_IMAGES

Required. This provides a mapping from the Kafka version to the corresponding Docker image containing a Kafka connect of that version. The required syntax is whitespace or comma separated <version>=<image> pairs. For example 2.7.0=quay.io/strimzi/kafka:0.24.0-kafka-2.7.0, 2.8.0=quay.io/strimzi/kafka:0.24.0-kafka-2.8.0. This is used when a KafkaConnect.spec.version property is specified but not the KafkaConnect.spec.image.

STRIMZI_KAFKA_CONNECT_S2I_IMAGES

Required. This provides a mapping from the Kafka version to the corresponding Docker image containing a Kafka connect of that version. The required syntax is whitespace or comma separated <version>=<image> pairs. For example 2.7.0=quay.io/strimzi/kafka:0.24.0-kafka-2.7.0, 2.8.0=quay.io/strimzi/kafka:0.24.0-kafka-2.8.0. This is used when a KafkaConnectS2I.spec.version property is specified but not the KafkaConnectS2I.spec.image.

STRIMZI_KAFKA_MIRROR_MAKER_IMAGES

Required. This provides a mapping from the Kafka version to the corresponding Docker image containing a Kafka mirror maker of that version. The required syntax is whitespace or comma separated <version>=<image> pairs. For example 2.7.0=quay.io/strimzi/kafka:0.24.0-kafka-2.7.0, 2.8.0=quay.io/strimzi/kafka:0.24.0-kafka-2.8.0. This is used when a KafkaMirrorMaker.spec.version property is specified but not the KafkaMirrorMaker.spec.image.

STRIMZI_DEFAULT_TOPIC_OPERATOR_IMAGE

Optional, default quay.io/strimzi/operator:0.24.0. The image name to use as the default when deploying the topic operator, if no image is specified as the Kafka.spec.entityOperator.topicOperator.image in Kafka resource.

STRIMZI_DEFAULT_USER_OPERATOR_IMAGE

Optional, default quay.io/strimzi/operator:0.24.0. The image name to use as the default when deploying the user operator, if no image is specified as the Kafka.spec.entityOperator.userOperator.image in the Kafka resource.

STRIMZI_DEFAULT_TLS_SIDECAR_ENTITY_OPERATOR_IMAGE

Optional, default quay.io/strimzi/kafka:0.24.0-kafka-2.8.0. The image name to use as the default when deploying the sidecar container which provides TLS support for the Entity Operator, if no image is specified as the Kafka.spec.entityOperator.tlsSidecar.image in the Kafka resource.

STRIMZI_IMAGE_PULL_POLICY

Optional. The ImagePullPolicy which will be applied to containers in all pods managed by Strimzi Cluster Operator. The valid values are Always, IfNotPresent, and Never. If not specified, the Kubernetes defaults will be used. Changing the policy will result in a rolling update of all your Kafka, Kafka Connect, and Kafka MirrorMaker clusters.

STRIMZI_IMAGE_PULL_SECRETS

Optional. A comma-separated list of Secret names. The secrets referenced here contain the credentials to the container registries where the container images are pulled from. The secrets are used in the imagePullSecrets field for all Pods created by the Cluster Operator. Changing this list results in a rolling update of all your Kafka, Kafka Connect, and Kafka MirrorMaker clusters.

STRIMZI_KUBERNETES_VERSION

Optional. Overrides the Kubernetes version information detected from the API server.

Example configuration for Kubernetes version override
env:
  - name: STRIMZI_KUBERNETES_VERSION
    value: |
           major=1
           minor=16
           gitVersion=v1.16.2
           gitCommit=c97fe5036ef3df2967d086711e6c0c405941e14b
           gitTreeState=clean
           buildDate=2019-10-15T19:09:08Z
           goVersion=go1.12.10
           compiler=gc
           platform=linux/amd64
KUBERNETES_SERVICE_DNS_DOMAIN

Optional. Overrides the default Kubernetes DNS domain name suffix.

By default, services assigned in the Kubernetes cluster have a DNS domain name that uses the default suffix cluster.local.

For example, for broker kafka-0:

<cluster-name>-kafka-0.<cluster-name>-kafka-brokers.<namespace>.svc.cluster.local

The DNS domain name is added to the Kafka broker certificates used for hostname verification.

If you are using a different DNS domain name suffix in your cluster, change the KUBERNETES_SERVICE_DNS_DOMAIN environment variable from the default to the one you are using in order to establish a connection with the Kafka brokers.

STRIMZI_CONNECT_BUILD_TIMEOUT_MS

Optional, default 300000 ms. The timeout for building new Kafka Connect images with additional connectots, in milliseconds. This value should be increased when using Strimzi to build container images containing many connectors or using a slow container registry.

STRIMZI_FEATURE_GATES

Optional. Enables or disables features and functionality controlled by feature gates. For more information about each feature gate, see Feature gates.

Feature gates

Strimzi operators support feature gates to enable or disable certain features and functionality. Enabling a feature gate changes the behavior of the relevant operator and introduces the feature to your Strimzi deployment.

Feature gates have a default state of either enabled or disabled. To modify a feature gate’s default state, use the STRIMZI_FEATURE_GATES environment variable in the operator’s configuration. You can modify multiple feature gates using this single environment variable.

Feature gates have three stages of maturity:

  • Alpha — typically disabled by default

  • Beta — typically enabled by default

  • General Availability (GA) — typically enabled by default

Alpha stage features might be experimental or unstable, subject to change, or not sufficiently tested for production use. Beta stage features are well tested and their functionality is not likely to change. GA stage features are stable and should not change in future. Alpha and beta stage features are removed if they do not prove to be useful.

Note
Feature gates might be removed when they reach GA. This means that the feature was incorporated into the Strimzi core features and can no longer be disabled.
Table 1. All feature gates and the Strimzi versions when they moved to alpha, beta, or GA
Feature gate Alpha Beta GA

ControlPlaneListener

0.23.0

-

-

ServiceAccountPatching

0.24.0

-

-

Configuring feature gates

You configure feature gates using the STRIMZI_FEATURE_GATES environment variable in the operator’s configuration. Specify a comma-separated list of feature gate names and prefixes. A + prefix enables the feature gate and a - prefix disables it.

Example feature gate configuration that enables FeatureGate1 and disables FeatureGate2
env:
  - name: STRIMZI_FEATURE_GATES
    value: +FeatureGate1,-FeatureGate2
Control plane listener feature gate

Use the ControlPlaneListener feature gate to change the communication paths used for inter-broker communications within your Kafka cluster.

The Kubernetes control plane manages the workloads running on the worker nodes. Services such as the Kubernetes API server and the controller manager run on the control plane. The Kubernetes data plane provides resources to containers, including CPU, memory, network, and storage.

In Strimzi, control plane traffic consists of controller connections that maintain the desired state of the Kafka cluster. Data plane traffic mainly consists of data replication between the leader broker and the follower brokers.

When the ControlPlaneListener feature gate is disabled, control plane and data plane traffic go through the same internal listener on port 9091. This was the default behavior before the feature gate was introduced.

When ControlPlaneListener is enabled, control plane traffic goes through a dedicated control plane listener on port 9090. Data plane traffic continues to use the internal listener on port 9091.

Using control plane listeners might improve performance because important controller connections, such as partition leadership changes, are not delayed by data replication across brokers.

Enabling the control plane listener feature gate

The ControlPlaneListener feature gate is in the alpha stage and has a default state of disabled. To enable it, specify +ControlPlaneListener in the STRIMZI_FEATURE_GATES environment variable in the Cluster Operator configuration.

This feature gate must be disabled when:

  • Upgrading from Strimzi 0.22 and earlier

  • Downgrading to Strimzi 0.22 and earlier

Note
The ControlPlaneListener feature gate was introduced in Strimzi 0.23.0 and is expected to remain in the alpha stage for a number of releases before moving to the beta stage.
Service Account patching feature gate

By default, the Cluster Operator does not update service accounts. To allow the Cluster Operator to apply updates, enable the ServiceAccountPatching feature gate.

Add +ServiceAccountPatching to the STRIMZI_FEATURE_GATES environment variable in the Cluster Operator configuration.

The feature gate is currently in the alpha phase and disabled by default. With the feature gate enabled, the Cluster Operator applies updates to service account configuration in every reconciliation. For example, you can change service account labels and annotations after the operands are already created.

Note
The ServiceAccountPatching feature gate was introduced in Strimzi 0.24.0 and is expected to remain in the alpha phase for a number of releases before it moves to the beta phase and is enabled by default.
Logging configuration by ConfigMap

The Cluster Operator’s logging is configured by the strimzi-cluster-operator ConfigMap.

A ConfigMap containing logging configuration is created when installing the Cluster Operator. This ConfigMap is described in the file install/cluster-operator/050-ConfigMap-strimzi-cluster-operator.yaml. You configure Cluster Operator logging by changing the data field log4j2.properties in this ConfigMap.

To update the logging configuration, you can edit the 050-ConfigMap-strimzi-cluster-operator.yaml file and then run the following command:

kubectl create -f install/cluster-operator/050-ConfigMap-strimzi-cluster-operator.yaml

Alternatively, edit the ConfigMap directly:

kubectl edit configmap strimzi-cluster-operator

To change the frequency of the reload interval, set a time in seconds in the monitorInterval option in the created ConfigMap.

If the ConfigMap is missing when the Cluster Operator is deployed, the default logging values are used.

If the ConfigMap is accidentally deleted after the Cluster Operator is deployed, the most recently loaded logging configuration is used. Create a new ConfigMap to load a new logging configuration.

Note
Do not remove the monitorInterval option from the ConfigMap.
Restricting Cluster Operator access with network policy

The Cluster Operator can run in the same namespace as the resources it manages, or in a separate namespace. By default, the STRIMZI_OPERATOR_NAMESPACE environment variable is configured to use the Kubernetes Downward API to find which namespace the Cluster Operator is running in. If the Cluster Operator is running in the same namespace as the resources, only local access is required, and allowed by Strimzi.

If the Cluster Operator is running in a separate namespace to the resources it manages, any namespace in the Kubernetes cluster is allowed access to the Cluster Operator unless network policy is configured. Use the optional STRIMZI_OPERATOR_NAMESPACE_LABELS environment variable to establish network policy for the Cluster Operator using namespace labels. By adding namespace labels, access to the Cluster Operator is restricted to the namespaces specified.

Network policy configured for the Cluster Operator deployment
#...
env:
  # ...
  - name: STRIMZI_OPERATOR_NAMESPACE_LABELS
    value: label1=value1,label2=value2
  #...
Periodic reconciliation

Although the Cluster Operator reacts to all notifications about the desired cluster resources received from the Kubernetes cluster, if the operator is not running, or if a notification is not received for any reason, the desired resources will get out of sync with the state of the running Kubernetes cluster.

In order to handle failovers properly, a periodic reconciliation process is executed by the Cluster Operator so that it can compare the state of the desired resources with the current cluster deployments in order to have a consistent state across all of them. You can set the time interval for the periodic reconciliations using the STRIMZI_FULL_RECONCILIATION_INTERVAL_MS variable.

5.1.2. Provisioning Role-Based Access Control (RBAC)

For the Cluster Operator to function it needs permission within the Kubernetes cluster to interact with resources such as Kafka, KafkaConnect, and so on, as well as the managed resources, such as ConfigMaps, Pods, Deployments, StatefulSets and Services. Such permission is described in terms of Kubernetes role-based access control (RBAC) resources:

  • ServiceAccount,

  • Role and ClusterRole,

  • RoleBinding and ClusterRoleBinding.

In addition to running under its own ServiceAccount with a ClusterRoleBinding, the Cluster Operator manages some RBAC resources for the components that need access to Kubernetes resources.

Kubernetes also includes privilege escalation protections that prevent components operating under one ServiceAccount from granting other ServiceAccounts privileges that the granting ServiceAccount does not have. Because the Cluster Operator must be able to create the ClusterRoleBindings, and RoleBindings needed by resources it manages, the Cluster Operator must also have those same privileges.

Delegated privileges

When the Cluster Operator deploys resources for a desired Kafka resource it also creates ServiceAccounts, RoleBindings, and ClusterRoleBindings, as follows:

  • The Kafka broker pods use a ServiceAccount called cluster-name-kafka

    • When the rack feature is used, the strimzi-cluster-name-kafka-init ClusterRoleBinding is used to grant this ServiceAccount access to the nodes within the cluster via a ClusterRole called strimzi-kafka-broker

    • When the rack feature is not used and the cluster is not exposed via nodeport, no binding is created

  • The ZooKeeper pods use a ServiceAccount called cluster-name-zookeeper

  • The Entity Operator pod uses a ServiceAccount called cluster-name-entity-operator

    • The Topic Operator produces Kubernetes events with status information, so the ServiceAccount is bound to a ClusterRole called strimzi-entity-operator which grants this access via the strimzi-entity-operator RoleBinding

  • The pods for KafkaConnect and KafkaConnectS2I resources use a ServiceAccount called cluster-name-cluster-connect

  • The pods for KafkaMirrorMaker use a ServiceAccount called cluster-name-mirror-maker

  • The pods for KafkaMirrorMaker2 use a ServiceAccount called cluster-name-mirrormaker2

  • The pods for KafkaBridge use a ServiceAccount called cluster-name-bridge

ServiceAccount

The Cluster Operator is best run using a ServiceAccount:

Example ServiceAccount for the Cluster Operator
apiVersion: v1
kind: ServiceAccount
metadata:
  name: strimzi-cluster-operator
  labels:
    app: strimzi

The Deployment of the operator then needs to specify this in its spec.template.spec.serviceAccountName:

Partial example of Deployment for the Cluster Operator
apiVersion: apps/v1
kind: Deployment
metadata:
  name: strimzi-cluster-operator
  labels:
    app: strimzi
spec:
  replicas: 1
  selector:
    matchLabels:
      name: strimzi-cluster-operator
      strimzi.io/kind: cluster-operator
  template:
      # ...

Note line 12, where the strimzi-cluster-operator ServiceAccount is specified as the serviceAccountName.

ClusterRoles

The Cluster Operator needs to operate using ClusterRoles that gives access to the necessary resources. Depending on the Kubernetes cluster setup, a cluster administrator might be needed to create the ClusterRoles.

Note
Cluster administrator rights are only needed for the creation of the ClusterRoles. The Cluster Operator will not run under the cluster admin account.

The ClusterRoles follow the principle of least privilege and contain only those privileges needed by the Cluster Operator to operate Kafka, Kafka Connect, and ZooKeeper clusters. The first set of assigned privileges allow the Cluster Operator to manage Kubernetes resources such as StatefulSets, Deployments, Pods, and ConfigMaps.

Cluster Operator uses ClusterRoles to grant permission at the namespace-scoped resources level and cluster-scoped resources level:

ClusterRole with namespaced resources for the Cluster Operator
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: strimzi-cluster-operator-namespaced
  labels:
    app: strimzi
rules:
  - apiGroups:
      - "rbac.authorization.k8s.io"
    resources:
      # The cluster operator needs to access and manage rolebindings to grant Strimzi components cluster permissions
      - rolebindings
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - "rbac.authorization.k8s.io"
    resources:
      # The cluster operator needs to access and manage roles to grant the entity operator permissions
      - roles
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - ""
    resources:
      # The cluster operator needs to access and delete pods, this is to allow it to monitor pod health and coordinate rolling updates
      - pods
      # The cluster operator needs to access and manage service accounts to grant Strimzi components cluster permissions
      - serviceaccounts
      # The cluster operator needs to access and manage config maps for Strimzi components configuration
      - configmaps
      # The cluster operator needs to access and manage services and endpoints to expose Strimzi components to network traffic
      - services
      - endpoints
      # The cluster operator needs to access and manage secrets to handle credentials
      - secrets
      # The cluster operator needs to access and manage persistent volume claims to bind them to Strimzi components for persistent data
      - persistentvolumeclaims
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - "kafka.strimzi.io"
    resources:
      # The cluster operator runs the KafkaAssemblyOperator, which needs to access and manage Kafka resources
      - kafkas
      - kafkas/status
      # The cluster operator runs the KafkaConnectAssemblyOperator, which needs to access and manage KafkaConnect resources
      - kafkaconnects
      - kafkaconnects/status
      # The cluster operator runs the KafkaConnectS2IAssemblyOperator, which needs to access and manage KafkaConnectS2I resources
      - kafkaconnects2is
      - kafkaconnects2is/status
      # The cluster operator runs the KafkaConnectorAssemblyOperator, which needs to access and manage KafkaConnector resources
      - kafkaconnectors
      - kafkaconnectors/status
      # The cluster operator runs the KafkaMirrorMakerAssemblyOperator, which needs to access and manage KafkaMirrorMaker resources
      - kafkamirrormakers
      - kafkamirrormakers/status
      # The cluster operator runs the KafkaBridgeAssemblyOperator, which needs to access and manage BridgeMaker resources
      - kafkabridges
      - kafkabridges/status
      # The cluster operator runs the KafkaMirrorMaker2AssemblyOperator, which needs to access and manage KafkaMirrorMaker2 resources
      - kafkamirrormaker2s
      - kafkamirrormaker2s/status
      # The cluster operator runs the KafkaRebalanceAssemblyOperator, which needs to access and manage KafkaRebalance resources
      - kafkarebalances
      - kafkarebalances/status
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      # The cluster operator needs the extensions api as the operator supports Kubernetes version 1.11+
      # apps/v1 was introduced in Kubernetes 1.14
      - "extensions"
    resources:
      # The cluster operator needs to access and manage deployments to run deployment based Strimzi components
      - deployments
      - deployments/scale
      # The cluster operator needs to access replica sets to manage Strimzi components and to determine error states
      - replicasets
      # The cluster operator needs to access and manage replication controllers to manage replicasets
      - replicationcontrollers
      # The cluster operator needs to access and manage network policies to lock down communication between Strimzi components
      - networkpolicies
      # The cluster operator needs to access and manage ingresses which allow external access to the services in a cluster
      - ingresses
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - "apps"
    resources:
      # The cluster operator needs to access and manage deployments to run deployment based Strimzi components
      - deployments
      - deployments/scale
      - deployments/status
      # The cluster operator needs to access and manage stateful sets to run stateful sets based Strimzi components
      - statefulsets
      # The cluster operator needs to access replica-sets to manage Strimzi components and to determine error states
      - replicasets
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - ""
    resources:
      # The cluster operator needs to be able to create events and delegate permissions to do so
      - events
    verbs:
      - create
  - apiGroups:
      # OpenShift S2I requirements
      - apps.openshift.io
    resources:
      - deploymentconfigs
      - deploymentconfigs/scale
      - deploymentconfigs/status
      - deploymentconfigs/finalizers
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      # OpenShift S2I requirements
      - build.openshift.io
    resources:
      - buildconfigs
      - buildconfigs/instantiate
      - builds
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      # OpenShift S2I requirements
      - image.openshift.io
    resources:
      - imagestreams
      - imagestreams/status
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - networking.k8s.io
    resources:
      # The cluster operator needs to access and manage network policies to lock down communication between Strimzi components
      - networkpolicies
      # The cluster operator needs to access and manage ingresses which allow external access to the services in a cluster
      - ingresses
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - route.openshift.io
    resources:
      # The cluster operator needs to access and manage routes to expose Strimzi components for external access
      - routes
      - routes/custom-host
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - policy
    resources:
      # The cluster operator needs to access and manage pod disruption budgets this limits the number of concurrent disruptions
      # that a Strimzi component experiences, allowing for higher availability
      - poddisruptionbudgets
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update

The second includes the permissions needed for cluster-scoped resources.

ClusterRole with cluster-scoped resources for the Cluster Operator
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: strimzi-cluster-operator-global
  labels:
    app: strimzi
rules:
  - apiGroups:
      - "rbac.authorization.k8s.io"
    resources:
      # The cluster operator needs to create and manage cluster role bindings in the case of an install where a user
      # has specified they want their cluster role bindings generated
      - clusterrolebindings
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update
  - apiGroups:
      - storage.k8s.io
    resources:
      # The cluster operator requires "get" permissions to view storage class details
      # This is because only a persistent volume of a supported storage class type can be resized
      - storageclasses
    verbs:
      - get
  - apiGroups:
      - ""
    resources:
      # The cluster operator requires "list" permissions to view all nodes in a cluster
      # The listing is used to determine the node addresses when NodePort access is configured
      # These addresses are then exposed in the custom resource states
      - nodes
    verbs:
      - list

The strimzi-kafka-broker ClusterRole represents the access needed by the init container in Kafka pods that is used for the rack feature. As described in the Delegated privileges section, this role is also needed by the Cluster Operator in order to be able to delegate this access.

ClusterRole for the Cluster Operator allowing it to delegate access to Kubernetes nodes to the Kafka broker pods
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: strimzi-kafka-broker
  labels:
    app: strimzi
rules:
  - apiGroups:
      - ""
    resources:
      # The Kafka Brokers require "get" permissions to view the node they are on
      # This information is used to generate a Rack ID that is used for High Availability configurations
      - nodes
    verbs:
      - get

The strimzi-topic-operator ClusterRole represents the access needed by the Topic Operator. As described in the Delegated privileges section, this role is also needed by the Cluster Operator in order to be able to delegate this access.

ClusterRole for the Cluster Operator allowing it to delegate access to events to the Topic Operator
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: strimzi-entity-operator
  labels:
    app: strimzi
rules:
  - apiGroups:
      - "kafka.strimzi.io"
    resources:
      # The entity operator runs the KafkaTopic assembly operator, which needs to access and manage KafkaTopic resources
      - kafkatopics
      - kafkatopics/status
      # The entity operator runs the KafkaUser assembly operator, which needs to access and manage KafkaUser resources
      - kafkausers
      - kafkausers/status
    verbs:
      - get
      - list
      - watch
      - create
      - patch
      - update
      - delete
  - apiGroups:
      - ""
    resources:
      - events
    verbs:
      # The entity operator needs to be able to create events
      - create
  - apiGroups:
      - ""
    resources:
      # The entity operator user-operator needs to access and manage secrets to store generated credentials
      - secrets
    verbs:
      - get
      - list
      - watch
      - create
      - delete
      - patch
      - update

The strimzi-kafka-client ClusterRole represents the access needed by the components based on Kafka clients which use the client rack-awareness. As described in the Delegated privileges section, this role is also needed by the Cluster Operator in order to be able to delegate this access.

ClusterRole for the Cluster Operator allowing it to delegate access to Kubernetes nodes to the Kafka client based pods
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: strimzi-kafka-client
  labels:
    app: strimzi
rules:
  - apiGroups:
      - ""
    resources:
      # The Kafka clients (Connect, Mirror Maker, etc.) require "get" permissions to view the node they are on
      # This information is used to generate a Rack ID (client.rack option) that is used for consuming from the closest
      # replicas when enabled
      - nodes
    verbs:
      - get
ClusterRoleBindings

The operator needs ClusterRoleBindings and RoleBindings which associates its ClusterRole with its ServiceAccount: ClusterRoleBindings are needed for ClusterRoles containing cluster-scoped resources.

Example ClusterRoleBinding for the Cluster Operator
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: strimzi-cluster-operator
  labels:
    app: strimzi
subjects:
  - kind: ServiceAccount
    name: strimzi-cluster-operator
    namespace: myproject
roleRef:
  kind: ClusterRole
  name: strimzi-cluster-operator-global
  apiGroup: rbac.authorization.k8s.io

ClusterRoleBindings are also needed for the ClusterRoles needed for delegation:

Example ClusterRoleBinding for the Cluster Operator for the Kafka broker rack-awareness
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: strimzi-cluster-operator-kafka-broker-delegation
  labels:
    app: strimzi
# The Kafka broker cluster role must be bound to the cluster operator service account so that it can delegate the cluster role to the Kafka brokers.
# This must be done to avoid escalating privileges which would be blocked by Kubernetes.
subjects:
  - kind: ServiceAccount
    name: strimzi-cluster-operator
    namespace: myproject
roleRef:
  kind: ClusterRole
  name: strimzi-kafka-broker
  apiGroup: rbac.authorization.k8s.io

and

Example ClusterRoleBinding for the Cluster Operator for the Kafka client rack-awareness
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: strimzi-cluster-operator-kafka-client-delegation
  labels:
    app: strimzi
# The Kafka clients cluster role must be bound to the cluster operator service account so that it can delegate the
# cluster role to the Kafka clients using it for consuming from closest replica.
# This must be done to avoid escalating privileges which would be blocked by Kubernetes.
subjects:
  - kind: ServiceAccount
    name: strimzi-cluster-operator
    namespace: myproject
roleRef:
  kind: ClusterRole
  name: strimzi-kafka-client
  apiGroup: rbac.authorization.k8s.io

ClusterRoles containing only namespaced resources are bound using RoleBindings only.

apiVersion: rbac.authorization.k8s.io/v1
kind: RoleBinding
metadata:
  name: strimzi-cluster-operator
  labels:
    app: strimzi
subjects:
  - kind: ServiceAccount
    name: strimzi-cluster-operator
    namespace: myproject
roleRef:
  kind: ClusterRole
  name: strimzi-cluster-operator-namespaced
  apiGroup: rbac.authorization.k8s.io
apiVersion: rbac.authorization.k8s.io/v1
kind: RoleBinding
metadata:
  name: strimzi-cluster-operator-entity-operator-delegation
  labels:
    app: strimzi
# The Entity Operator cluster role must be bound to the cluster operator service account so that it can delegate the cluster role to the Entity Operator.
# This must be done to avoid escalating privileges which would be blocked by Kubernetes.
subjects:
  - kind: ServiceAccount
    name: strimzi-cluster-operator
    namespace: myproject
roleRef:
  kind: ClusterRole
  name: strimzi-entity-operator
  apiGroup: rbac.authorization.k8s.io

5.2. Using the Topic Operator

When you create, modify or delete a topic using the KafkaTopic resource, the Topic Operator ensures those changes are reflected in the Kafka cluster.

The Deploying and Upgrading Strimzi guide provides instructions to deploy the Topic Operator:

5.2.1. Kafka topic resource

The KafkaTopic resource is used to configure topics, including the number of partitions and replicas.

The full schema for KafkaTopic is described in KafkaTopic schema reference.

Identifying a Kafka cluster for topic handling

A KafkaTopic resource includes a label that defines the appropriate name of the Kafka cluster (derived from the name of the Kafka resource) to which it belongs.

For example:

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaTopic
metadata:
  name: topic-name-1
  labels:
    strimzi.io/cluster: my-cluster

The label is used by the Topic Operator to identify the KafkaTopic resource and create a new topic, and also in subsequent handling of the topic.

If the label does not match the Kafka cluster, the Topic Operator cannot identify the KafkaTopic and the topic is not created.

Kafka topic usage recommendations

When working with topics, be consistent. Always operate on either KafkaTopic resources or topics directly in Kubernetes. Avoid routinely switching between both methods for a given topic.

Use topic names that reflect the nature of the topic, and remember that names cannot be changed later.

If creating a topic in Kafka, use a name that is a valid Kubernetes resource name, otherwise the Topic Operator will need to create the corresponding KafkaTopic with a name that conforms to the Kubernetes rules.

Note
Recommendations for identifiers and names in Kubernetes are outlined in Identifiers and Names in Kubernetes community article.
Kafka topic naming conventions

Kafka and Kubernetes impose their own validation rules for the naming of topics in Kafka and KafkaTopic.metadata.name respectively. There are valid names for each which are invalid in the other.

Using the spec.topicName property, it is possible to create a valid topic in Kafka with a name that would be invalid for the Kafka topic in Kubernetes.

The spec.topicName property inherits Kafka naming validation rules:

  • The name must not be longer than 249 characters.

  • Valid characters for Kafka topics are ASCII alphanumerics, ., _, and -.

  • The name cannot be . or .., though . can be used in a name, such as exampleTopic. or .exampleTopic.

spec.topicName must not be changed.

For example:

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaTopic
metadata:
  name: topic-name-1
spec:
  topicName: topicName-1 (1)
  # ...
  1. Upper case is invalid in Kubernetes.

cannot be changed to:

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaTopic
metadata:
  name: topic-name-1
spec:
  topicName: name-2
  # ...
Note

Some Kafka client applications, such as Kafka Streams, can create topics in Kafka programmatically. If those topics have names that are invalid Kubernetes resource names, the Topic Operator gives them a valid metadata.name based on the Kafka name. Invalid characters are replaced and a hash is appended to the name. For example:

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaTopic
metadata:
  name: mytopic---c55e57fe2546a33f9e603caf57165db4072e827e
spec:
  topicName: myTopic
  # ...

5.2.2. Topic Operator topic store

The Topic Operator uses Kafka to store topic metadata describing topic configuration as key-value pairs. The topic store is based on the Kafka Streams key-value mechanism, which uses Kafka topics to persist the state.

Topic metadata is cached in-memory and accessed locally within the Topic Operator. Updates from operations applied to the local in-memory cache are persisted to a backup topic store on disk. The topic store is continually synchronized with updates from Kafka topics or Kubernetes KafkaTopic custom resources. Operations are handled rapidly with the topic store set up this way, but should the in-memory cache crash it is automatically repopulated from the persistent storage.

Internal topic store topics

Internal topics support the handling of topic metadata in the topic store.

__strimzi_store_topic

Input topic for storing the topic metadata

__strimzi-topic-operator-kstreams-topic-store-changelog

Retains a log of compacted topic store values

Warning
Do not delete these topics, as they are essential to the running of the Topic Operator.
Migrating topic metadata from ZooKeeper

In previous releases of Strimzi, topic metadata was stored in ZooKeeper. The new process removes this requirement, bringing the metadata into the Kafka cluster, and under the control of the Topic Operator.

When upgrading to Strimzi 0.24.0, the transition to Topic Operator control of the topic store is seamless. Metadata is found and migrated from ZooKeeper, and the old store is deleted.

Downgrading to a Strimzi version that uses ZooKeeper to store topic metadata

If you are reverting back to a version of Strimzi earlier than 0.22, which uses ZooKeeper for the storage of topic metadata, you still downgrade your Cluster Operator to the previous version, then downgrade Kafka brokers and client applications to the previous Kafka version as standard.

However, you must also delete the topics that were created for the topic store using a kafka-admin command, specifying the bootstrap address of the Kafka cluster. For example:

kubectl run kafka-admin -ti --image=quay.io/strimzi/kafka:0.24.0-kafka-2.8.0 --rm=true --restart=Never -- ./bin/kafka-topics.sh --bootstrap-server localhost:9092 --topic __strimzi-topic-operator-kstreams-topic-store-changelog --delete && ./bin/kafka-topics.sh --bootstrap-server localhost:9092 --topic __strimzi_store_topic --delete

The command must correspond to the type of listener and authentication used to access the Kafka cluster.

The Topic Operator will reconstruct the ZooKeeper topic metadata from the state of the topics in Kafka.

Topic Operator topic replication and scaling

The recommended configuration for topics managed by the Topic Operator is a topic replication factor of 3, and a minimum of 2 in-sync replicas.

apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaTopic
metadata:
  name: my-topic
  labels:
    strimzi.io/cluster: my-cluster
spec:
  partitions: 1 (1)
  replicas: 3 (2)
  config:
    min.insync.replicas=2 (3)
  #...
  1. The number of partitions for the topic. Generally, 1 partition is sufficient.

  2. The number of replica topic partitions. Currently, this cannot be changed in the KafkaTopic resource, but it can be changed using the kafka-reassign-partitions.sh tool.

  3. The minimum number of replica partitions that a message must be successfully written to, or an exception is raised.

Note
In-sync replicas are used in conjunction with the acks configuration for producer applications. The acks configuration determines the number of follower partitions a message must be replicated to before the message is acknowledged as successfully received. The Topic Operator runs with acks=all, whereby messages must be acknowledged by all in-sync replicas.

When scaling Kafka clusters by adding or removing brokers, replication factor configuration is not changed and replicas are not reassigned automatically. However, you can use the kafka-reassign-partitions.sh tool to change the replication factor, and manually reassign replicas to brokers.

Alternatively, though the integration of Cruise Control for Strimzi cannot change the replication factor for topics, the optimization proposals it generates for rebalancing Kafka include commands that transfer partition replicas and change partition leadership.

Handling changes to topics

A fundamental problem that the Topic Operator needs to solve is that there is no single source of truth: both the KafkaTopic resource and the Kafka topic can be modified independently of the Topic Operator. Complicating this, the Topic Operator might not always be able to observe changes at each end in real time. For example, when the Topic Operator is down.

To resolve this, the Topic Operator maintains information about each topic in the topic store. When a change happens in the Kafka cluster or Kubernetes, it looks at both the state of the other system and the topic store in order to determine what needs to change to keep everything in sync. The same thing happens whenever the Topic Operator starts, and periodically while it is running.

For example, suppose the Topic Operator is not running, and a KafkaTopic called my-topic is created. When the Topic Operator starts, the topic store does not contain information on my-topic, so it can infer that the KafkaTopic was created after it was last running. The Topic Operator creates the topic corresponding to my-topic, and also stores metadata for my-topic in the topic store.

If you update Kafka topic configuration or apply a change through the KafkaTopic custom resource, the topic store is updated after the Kafka cluster is reconciled.

The topic store also allows the Topic Operator to manage scenarios where the topic configuration is changed in Kafka topics and updated through Kubernetes KafkaTopic custom resources, as long as the changes are not incompatible. For example, it is possible to make changes to the same topic config key, but to different values. For incompatible changes, the Kafka configuration takes priority, and the KafkaTopic is updated accordingly.

Note
You can also use the KafkaTopic resource to delete topics using a kubectl delete -f KAFKA-TOPIC-CONFIG-FILE command. To be able to do this, delete.topic.enable must be set to true (default) in the spec.kafka.config of the Kafka resource.

5.2.3. Configuring a Kafka topic

Use the properties of the KafkaTopic resource to configure a Kafka topic.

You can use kubectl apply to create or modify topics, and kubectl delete to delete existing topics.

For example:

  • kubectl apply -f <topic-config-file>

  • kubectl delete KafkaTopic <topic-name>

This procedure shows how to create a topic with 10 partitions and 2 replicas.

Before you start

It is important that you consider the following before making your changes:

  • Kafka does not support making the following changes through the KafkaTopic resource:

    • Changing topic names using spec.topicName

    • Decreasing partition size using spec.partitions

  • You cannot use spec.replicas to change the number of replicas that were initially specified.

  • Increasing spec.partitions for topics with keys will change how records are partitioned, which can be particularly problematic when the topic uses semantic partitioning.

Prerequisites
Procedure
  1. Prepare a file containing the KafkaTopic to be created.

    An example KafkaTopic
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaTopic
    metadata:
      name: orders
      labels:
        strimzi.io/cluster: my-cluster
    spec:
      partitions: 10
      replicas: 2
    Tip
    When modifying a topic, you can get the current version of the resource using kubectl get kafkatopic orders -o yaml.
  2. Create the KafkaTopic resource in Kubernetes.

    kubectl apply -f TOPIC-CONFIG-FILE

5.2.4. Configuring the Topic Operator with resource requests and limits

You can allocate resources, such as CPU and memory, to the Topic Operator and set a limit on the amount of resources it can consume.

Prerequisites
  • The Cluster Operator is running.

Procedure
  1. Update the Kafka cluster configuration in an editor, as required:

    kubectl edit kafka MY-CLUSTER
  2. In the spec.entityOperator.topicOperator.resources property in the Kafka resource, set the resource requests and limits for the Topic Operator.

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      # Kafka and ZooKeeper sections...
      entityOperator:
        topicOperator:
          resources:
            requests:
              cpu: "1"
              memory: 500Mi
            limits:
              cpu: "1"
              memory: 500Mi
  3. Apply the new configuration to create or update the resource.

    kubectl apply -f KAFKA-CONFIG-FILE

5.3. Using the User Operator

When you create, modify or delete a user using the KafkaUser resource, the User Operator ensures those changes are reflected in the Kafka cluster.

The Deploying and Upgrading Strimzi guide provides instructions to deploy the User Operator:

For more information about the schema, see KafkaUser schema reference.

Authenticating and authorizing access to Kafka

Use KafkaUser to enable the authentication and authorization mechanisms that a specific client uses to access Kafka.

For more information on using KafkUser to manage users and secure access to Kafka brokers, see Securing access to Kafka brokers.

5.3.1. Configuring the User Operator with resource requests and limits

You can allocate resources, such as CPU and memory, to the User Operator and set a limit on the amount of resources it can consume.

Prerequisites
  • The Cluster Operator is running.

Procedure
  1. Update the Kafka cluster configuration in an editor, as required:

    kubectl edit kafka MY-CLUSTER
  2. In the spec.entityOperator.userOperator.resources property in the Kafka resource, set the resource requests and limits for the User Operator.

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    spec:
      # Kafka and ZooKeeper sections...
      entityOperator:
        userOperator:
          resources:
            requests:
              cpu: "1"
              memory: 500Mi
            limits:
              cpu: "1"
              memory: 500Mi

    Save the file and exit the editor. The Cluster Operator applies the changes automatically.

5.4. Monitoring operators using Prometheus metrics

Strimzi operators expose Prometheus metrics. The metrics are automatically enabled and contain information about:

  • Number of reconciliations

  • Number of Custom Resources the operator is processing

  • Duration of reconciliations

  • JVM metrics from the operators

Additionally, we provide an example Grafana dashboard.

For more information about Prometheus, see the Introducing Metrics to Kafka in the Deploying and Upgrading Strimzi guide.

6. Kafka Bridge

This chapter provides an overview of the Strimzi Kafka Bridge and helps you get started using its REST API to interact with Strimzi.

6.1. Kafka Bridge overview

You can use the Strimzi Kafka Bridge as an interface to make specific types of HTTP requests to the Kafka cluster.

6.1.1. Kafka Bridge interface

The Kafka Bridge provides a RESTful interface that allows HTTP-based clients to interact with a Kafka cluster.  It offers the advantages of a web API connection to Strimzi, without the need for client applications to interpret the Kafka protocol.

The API has two main resources — consumers and topics — that are exposed and made accessible through endpoints to interact with consumers and producers in your Kafka cluster. The resources relate only to the Kafka Bridge, not the consumers and producers connected directly to Kafka.

HTTP requests

The Kafka Bridge supports HTTP requests to a Kafka cluster, with methods to:

  • Send messages to a topic.

  • Retrieve messages from topics.

  • Retrieve a list of partitions for a topic.

  • Create and delete consumers.

  • Subscribe consumers to topics, so that they start receiving messages from those topics.

  • Retrieve a list of topics that a consumer is subscribed to.

  • Unsubscribe consumers from topics.

  • Assign partitions to consumers.

  • Commit a list of consumer offsets.

  • Seek on a partition, so that a consumer starts receiving messages from the first or last offset position, or a given offset position.

The methods provide JSON responses and HTTP response code error handling. Messages can be sent in JSON or binary formats.

Clients can produce and consume messages without the requirement to use the native Kafka protocol.

Additional resources

6.1.2. Supported clients for the Kafka Bridge

You can use the Kafka Bridge to integrate both internal and external HTTP client applications with your Kafka cluster.

Internal clients

Internal clients are container-based HTTP clients running in the same Kubernetes cluster as the Kafka Bridge itself. Internal clients can access the Kafka Bridge on the host and port defined in the KafkaBridge custom resource.

External clients

External clients are HTTP clients running outside the Kubernetes cluster in which the Kafka Bridge is deployed and running. External clients can access the Kafka Bridge through an OpenShift Route, a loadbalancer service, or using an Ingress.

HTTP internal and external client integration

Internal and external HTTP producers and consumers exchange data with the Kafka brokers through the Kafka Bridge

6.1.3. Securing the Kafka Bridge

Strimzi does not currently provide any encryption, authentication, or authorization for the Kafka Bridge. This means that requests sent from external clients to the Kafka Bridge are:

  • Not encrypted, and must use HTTP rather than HTTPS

  • Sent without authentication

However, you can secure the Kafka Bridge using other methods, such as:

  • Kubernetes Network Policies that define which pods can access the Kafka Bridge.

  • Reverse proxies with authentication or authorization, for example, OAuth2 proxies.

  • API Gateways.

  • Ingress or OpenShift Routes with TLS termination.

The Kafka Bridge supports TLS encryption and TLS and SASL authentication when connecting to the Kafka Brokers. Within your Kubernetes cluster, you can configure:

  • TLS or SASL-based authentication between the Kafka Bridge and your Kafka cluster

  • A TLS-encrypted connection between the Kafka Bridge and your Kafka cluster.

For more information, see Configuring the Kafka Bridge.

You can use ACLs in Kafka brokers to restrict the topics that can be consumed and produced using the Kafka Bridge.

6.1.4. Accessing the Kafka Bridge outside of Kubernetes

After deployment, the Strimzi Kafka Bridge can only be accessed by applications running in the same Kubernetes cluster. These applications use the kafka-bridge-name-bridge-service Service to access the API.

If you want to make the Kafka Bridge accessible to applications running outside of the Kubernetes cluster, you can expose it manually by using one of the following features:

  • Services of types LoadBalancer or NodePort

  • Ingress resources

  • OpenShift Routes

If you decide to create Services, use the following labels in the selector to configure the pods to which the service will route the traffic:

  # ...
  selector:
    strimzi.io/cluster: kafka-bridge-name (1)
    strimzi.io/kind: KafkaBridge
  #...
  1. Name of the Kafka Bridge custom resource in your Kubernetes cluster.

6.1.5. Requests to the Kafka Bridge

Specify data formats and HTTP headers to ensure valid requests are submitted to the Kafka Bridge.

Content Type headers

API request and response bodies are always encoded as JSON.

  • When performing consumer operations, POST requests must provide the following Content-Type header if there is a non-empty body:

    Content-Type: application/vnd.kafka.v2+json
  • When performing producer operations, POST requests must provide Content-Type headers specifying the embedded data format of the messages produced. This can be either json or binary.

    Embedded data format Content-Type header

    JSON

    Content-Type: application/vnd.kafka.json.v2+json

    Binary

    Content-Type: application/vnd.kafka.binary.v2+json

The embedded data format is set per consumer, as described in the next section.

The Content-Type must not be set if the POST request has an empty body. An empty body can be used to create a consumer with the default values.

Embedded data format

The embedded data format is the format of the Kafka messages that are transmitted, over HTTP, from a producer to a consumer using the Kafka Bridge. Two embedded data formats are supported: JSON and binary.

When creating a consumer using the /consumers/groupid endpoint, the POST request body must specify an embedded data format of either JSON or binary. This is specified in the format field, for example:

{
  "name": "my-consumer",
  "format": "binary", (1)
...
}
  1. A binary embedded data format.

The embedded data format specified when creating a consumer must match the data format of the Kafka messages it will consume.

If you choose to specify a binary embedded data format, subsequent producer requests must provide the binary data in the request body as Base64-encoded strings. For example, when sending messages using the /topics/topicname endpoint, records.value must be encoded in Base64:

{
  "records": [
    {
      "key": "my-key",
      "value": "ZWR3YXJkdGhldGhyZWVsZWdnZWRjYXQ="
    },
  ]
}

Producer requests must also provide a Content-Type header that corresponds to the embedded data format, for example, Content-Type: application/vnd.kafka.binary.v2+json.

Message format

When sending messages using the /topics endpoint, you enter the message payload in the request body, in the records parameter.

The records parameter can contain any of these optional fields:

  • Message headers

  • Message key

  • Message value

  • Destination partition

Example POST request to /topics
curl -X POST \
  http://localhost:8080/topics/my-topic \
  -H 'content-type: application/vnd.kafka.json.v2+json' \
  -d '{
    "records": [
        {
            "key": "my-key",
            "value": "sales-lead-0001"
            "partition": 2
            "headers": [
              {
                "key": "key1",
                "value": "QXBhY2hlIEthZmthIGlzIHRoZSBib21iIQ==" (1)
              }
            ]
        },
    ]
}'
  1. The header value in binary format and encoded as Base64.

Accept headers

After creating a consumer, all subsequent GET requests must provide an Accept header in the following format:

Accept: application/vnd.kafka.EMBEDDED-DATA-FORMAT.v2+json

The EMBEDDED-DATA-FORMAT is either json or binary.

For example, when retrieving records for a subscribed consumer using an embedded data format of JSON, include this Accept header:

Accept: application/vnd.kafka.json.v2+json

6.1.6. CORS

Cross-Origin Resource Sharing (CORS) allows you to specify allowed methods and originating URLs for accessing the Kafka cluster in your Kafka Bridge HTTP configuration.

Example CORS configuration for Kafka Bridge
# ...
cors:
  allowedOrigins: "https://strimzi.io"
  allowedMethods: "GET,POST,PUT,DELETE,OPTIONS,PATCH"
  # ...

CORS allows for simple and preflighted requests between origin sources on different domains.

Simple requests are suitable for standard requests using GET, HEAD, POST methods.

A preflighted request sends a HTTP OPTIONS request as an initial check that the actual request is safe to send. On confirmation, the actual request is sent. Preflight requests are suitable for methods that require greater safeguards, such as PUT and DELETE, and use non-standard headers.

All requests require an Origin value in their header, which is the source of the HTTP request.

Simple request

For example, this simple request header specifies the origin as https://strimzi.io.

Origin: https://strimzi.io

The header information is added to the request.

curl -v -X GET HTTP-ADDRESS/bridge-consumer/records \
-H 'Origin: https://strimzi.io'\
-H 'content-type: application/vnd.kafka.v2+json'

In the response from the Kafka Bridge, an Access-Control-Allow-Origin header is returned.

HTTP/1.1 200 OK
Access-Control-Allow-Origin: * (1)
  1. Returning an asterisk (*) shows the resource can be accessed by any domain.

Preflighted request

An initial preflight request is sent to Kafka Bridge using an OPTIONS method. The HTTP OPTIONS request sends header information to check that Kafka Bridge will allow the actual request.

Here the preflight request checks that a POST request is valid from https://strimzi.io.

OPTIONS /my-group/instances/my-user/subscription HTTP/1.1
Origin: https://strimzi.io
Access-Control-Request-Method: POST (1)
Access-Control-Request-Headers: Content-Type (2)
  1. Kafka Bridge is alerted that the actual request is a POST request.

  2. The actual request will be sent with a Content-Type header.

OPTIONS is added to the header information of the preflight request.

curl -v -X OPTIONS -H 'Origin: https://strimzi.io' \
-H 'Access-Control-Request-Method: POST' \
-H 'content-type: application/vnd.kafka.v2+json'

Kafka Bridge responds to the initial request to confirm that the request will be accepted. The response header returns allowed origins, methods and headers.

HTTP/1.1 200 OK
Access-Control-Allow-Origin: https://strimzi.io
Access-Control-Allow-Methods: GET,POST,PUT,DELETE,OPTIONS,PATCH
Access-Control-Allow-Headers: content-type

If the origin or method is rejected, an error message is returned.

The actual request does not require Access-Control-Request-Method header, as it was confirmed in the preflight request, but it does require the origin header.

curl -v -X POST HTTP-ADDRESS/topics/bridge-topic \
-H 'Origin: https://strimzi.io' \
-H 'content-type: application/vnd.kafka.v2+json'

The response shows the originating URL is allowed.

HTTP/1.1 200 OK
Access-Control-Allow-Origin: https://strimzi.io
Additional resources

Fetch CORS specification

6.1.7. Kafka Bridge API resources

For the full list of REST API endpoints and descriptions, including example requests and responses, see the Kafka Bridge API reference.

6.1.8. Kafka Bridge deployment

You deploy the Kafka Bridge into your Kubernetes cluster by using the Cluster Operator.

After the Kafka Bridge is deployed, the Cluster Operator creates Kafka Bridge objects in your Kubernetes cluster. Objects include the deployment, service, and pod, each named after the name given in the custom resource for the Kafka Bridge.

Additional resources

6.2. Kafka Bridge quickstart

Use this quickstart to try out the Strimzi Kafka Bridge in your local development environment. You will learn how to:

  • Deploy the Kafka Bridge to your Kubernetes cluster

  • Expose the Kafka Bridge service to your local machine by using port-forwarding

  • Produce messages to topics and partitions in your Kafka cluster

  • Create a Kafka Bridge consumer

  • Perform basic consumer operations, such as subscribing the consumer to topics and retrieving the messages that you produced

In this quickstart, HTTP requests are formatted as curl commands that you can copy and paste to your terminal. Access to a Kubernetes cluster is required.

Ensure you have the prerequisites and then follow the tasks in the order provided in this chapter.

About data formats

In this quickstart, you will produce and consume messages in JSON format, not binary. For more information on the data formats and HTTP headers used in the example requests, see Requests to the Kafka Bridge.

Prerequisites for the quickstart
  • Cluster administrator access to a local or remote Kubernetes cluster.

  • Strimzi is installed.

  • A running Kafka cluster, deployed by the Cluster Operator, in a Kubernetes namespace.

  • The Entity Operator is deployed and running as part of the Kafka cluster.

6.2.1. Deploying the Kafka Bridge to your Kubernetes cluster

Strimzi includes a YAML example that specifies the configuration of the Strimzi Kafka Bridge. Make some minimal changes to this file and then deploy an instance of the Kafka Bridge to your Kubernetes cluster.

Procedure
  1. Edit the examples/bridge/kafka-bridge.yaml file.

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaBridge
    metadata:
      name: quickstart (1)
    spec:
      replicas: 1
      bootstrapServers: <cluster-name>-kafka-bootstrap:9092 (2)
      http:
        port: 8080
    1. When the Kafka Bridge is deployed, -bridge is appended to the name of the deployment and other related resources. In this example, the Kafka Bridge deployment is named quickstart-bridge and the accompanying Kafka Bridge service is named quickstart-bridge-service.

    2. In the bootstrapServers property, enter the name of the Kafka cluster as the <cluster-name>.

  2. Deploy the Kafka Bridge to your Kubernetes cluster:

    kubectl apply -f examples/bridge/kafka-bridge.yaml

    A quickstart-bridge deployment, service, and other related resources are created in your Kubernetes cluster.

  3. Verify that the Kafka Bridge was successfully deployed:

    kubectl get deployments
    NAME                             READY   UP-TO-DATE   AVAILABLE   AGE
    quickstart-bridge                  1/1     1            1          34m
    my-cluster-connect                 1/1     1            1          24h
    my-cluster-entity-operator         1/1     1            1          24h
    #...
What to do next

After deploying the Kafka Bridge to your Kubernetes cluster, expose the Kafka Bridge service to your local machine.

Additional resources

6.2.2. Exposing the Kafka Bridge service to your local machine

Next, use port forwarding to expose the Strimzi Kafka Bridge service to your local machine on http://localhost:8080.

Note
Port forwarding is only suitable for development and testing purposes.
Procedure
  1. List the names of the pods in your Kubernetes cluster:

    kubectl get pods -o name
    
    pod/kafka-consumer
    # ...
    pod/quickstart-bridge-589d78784d-9jcnr
    pod/strimzi-cluster-operator-76bcf9bc76-8dnfm
  2. Connect to the quickstart-bridge pod on port 8080:

    kubectl port-forward pod/quickstart-bridge-589d78784d-9jcnr 8080:8080 &
    Note
    If port 8080 on your local machine is already in use, use an alternative HTTP port, such as 8008.

API requests are now forwarded from port 8080 on your local machine to port 8080 in the Kafka Bridge pod.

6.2.3. Producing messages to topics and partitions

Next, produce messages to topics in JSON format by using the topics endpoint. You can specify destination partitions for messages in the request body, as shown here. The partitions endpoint provides an alternative method for specifying a single destination partition for all messages as a path parameter.

Procedure
  1. In a text editor, create a YAML definition for a Kafka topic with three partitions.

    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaTopic
    metadata:
      name: bridge-quickstart-topic
      labels:
        strimzi.io/cluster: <kafka-cluster-name> (1)
    spec:
      partitions: 3 (2)
      replicas: 1
      config:
        retention.ms: 7200000
        segment.bytes: 1073741824
    1. The name of the Kafka cluster in which the Kafka Bridge is deployed.

    2. The number of partitions for the topic.

  2. Save the file to the examples/topic directory as bridge-quickstart-topic.yaml.

  3. Create the topic in your Kubernetes cluster:

    kubectl apply -f examples/topic/bridge-quickstart-topic.yaml
  4. Using the Kafka Bridge, produce three messages to the topic you created:

    curl -X POST \
      http://localhost:8080/topics/bridge-quickstart-topic \
      -H 'content-type: application/vnd.kafka.json.v2+json' \
      -d '{
        "records": [
            {
                "key": "my-key",
                "value": "sales-lead-0001"
            },
            {
                "value": "sales-lead-0002",
                "partition": 2
            },
            {
                "value": "sales-lead-0003"
            }
        ]
    }'
    • sales-lead-0001 is sent to a partition based on the hash of the key.

    • sales-lead-0002 is sent directly to partition 2.

    • sales-lead-0003 is sent to a partition in the bridge-quickstart-topic topic using a round-robin method.

  5. If the request is successful, the Kafka Bridge returns an offsets array, along with a 200 code and a content-type header of application/vnd.kafka.v2+json. For each message, the offsets array describes:

    • The partition that the message was sent to

    • The current message offset of the partition

      Example response
      #...
      {
        "offsets":[
          {
            "partition":0,
            "offset":0
          },
          {
            "partition":2,
            "offset":0
          },
          {
            "partition":0,
            "offset":1
          }
        ]
      }
What to do next

After producing messages to topics and partitions, create a Kafka Bridge consumer.

Additional resources

6.2.4. Creating a Kafka Bridge consumer

Before you can perform any consumer operations in the Kafka cluster, you must first create a consumer by using the consumers endpoint. The consumer is referred to as a Kafka Bridge consumer.

Procedure
  1. Create a Kafka Bridge consumer in a new consumer group named bridge-quickstart-consumer-group:

    curl -X POST http://localhost:8080/consumers/bridge-quickstart-consumer-group \
      -H 'content-type: application/vnd.kafka.v2+json' \
      -d '{
        "name": "bridge-quickstart-consumer",
        "auto.offset.reset": "earliest",
        "format": "json",
        "enable.auto.commit": false,
        "fetch.min.bytes": 512,
        "consumer.request.timeout.ms": 30000
      }'
    • The consumer is named bridge-quickstart-consumer and the embedded data format is set as json.

    • Some basic configuration settings are defined.

    • The consumer will not commit offsets to the log automatically because the enable.auto.commit setting is false. You will commit the offsets manually later in this quickstart.

      If the request is successful, the Kafka Bridge returns the consumer ID (instance_id) and base URL (base_uri) in the response body, along with a 200 code.

      Example response
      #...
      {
        "instance_id": "bridge-quickstart-consumer",
        "base_uri":"http://<bridge-name>-bridge-service:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer"
      }
  2. Copy the base URL (base_uri) to use in the other consumer operations in this quickstart.

What to do next

Now that you have created a Kafka Bridge consumer, you can subscribe it to topics.

Additional resources

6.2.5. Subscribing a Kafka Bridge consumer to topics

After you have created a Kafka Bridge consumer, subscribe it to one or more topics by using the subscription endpoint. Once subscribed, the consumer starts receiving all messages that are produced to the topic.

Procedure
  • Subscribe the consumer to the bridge-quickstart-topic topic that you created earlier, in Producing messages to topics and partitions:

    curl -X POST http://localhost:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer/subscription \
      -H 'content-type: application/vnd.kafka.v2+json' \
      -d '{
        "topics": [
            "bridge-quickstart-topic"
        ]
    }'

    The topics array can contain a single topic (as shown here) or multiple topics. If you want to subscribe the consumer to multiple topics that match a regular expression, you can use the topic_pattern string instead of the topics array.

    If the request is successful, the Kafka Bridge returns a 204 (No Content) code only.

What to do next

After subscribing a Kafka Bridge consumer to topics, you can retrieve messages from the consumer.

Additional resources

6.2.6. Retrieving the latest messages from a Kafka Bridge consumer

Next, retrieve the latest messages from the Kafka Bridge consumer by requesting data from the records endpoint. In production, HTTP clients can call this endpoint repeatedly (in a loop).

Procedure
  1. Produce additional messages to the Kafka Bridge consumer, as described in Producing messages to topics and partitions.

  2. Submit a GET request to the records endpoint:

    curl -X GET http://localhost:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer/records \
      -H 'accept: application/vnd.kafka.json.v2+json'

    After creating and subscribing to a Kafka Bridge consumer, a first GET request will return an empty response because the poll operation starts a rebalancing process to assign partitions.

  3. Repeat step two to retrieve messages from the Kafka Bridge consumer.

    The Kafka Bridge returns an array of messages — describing the topic name, key, value, partition, and offset — in the response body, along with a 200 code. Messages are retrieved from the latest offset by default.

    HTTP/1.1 200 OK
    content-type: application/vnd.kafka.json.v2+json
    #...
    [
      {
        "topic":"bridge-quickstart-topic",
        "key":"my-key",
        "value":"sales-lead-0001",
        "partition":0,
        "offset":0
      },
      {
        "topic":"bridge-quickstart-topic",
        "key":null,
        "value":"sales-lead-0003",
        "partition":0,
        "offset":1
      },
    #...
    Note
    If an empty response is returned, produce more records to the consumer as described in Producing messages to topics and partitions, and then try retrieving messages again.
What to do next

After retrieving messages from a Kafka Bridge consumer, try committing offsets to the log.

Additional resources

6.2.7. Commiting offsets to the log

Next, use the offsets endpoint to manually commit offsets to the log for all messages received by the Kafka Bridge consumer. This is required because the Kafka Bridge consumer that you created earlier, in Creating a Kafka Bridge consumer, was configured with the enable.auto.commit setting as false.

Procedure
  • Commit offsets to the log for the bridge-quickstart-consumer:

    curl -X POST http://localhost:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer/offsets

    Because no request body is submitted, offsets are committed for all the records that have been received by the consumer. Alternatively, the request body can contain an array (OffsetCommitSeekList) that specifies the topics and partitions that you want to commit offsets for.

    If the request is successful, the Kafka Bridge returns a 204 code only.

What to do next

After committing offsets to the log, try out the endpoints for seeking to offsets.

Additional resources

6.2.8. Seeking to offsets for a partition

Next, use the positions endpoints to configure the Kafka Bridge consumer to retrieve messages for a partition from a specific offset, and then from the latest offset. This is referred to in Apache Kafka as a seek operation.

Procedure
  1. Seek to a specific offset for partition 0 of the quickstart-bridge-topic topic:

    curl -X POST http://localhost:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer/positions \
      -H 'content-type: application/vnd.kafka.v2+json' \
      -d '{
        "offsets": [
            {
                "topic": "bridge-quickstart-topic",
                "partition": 0,
                "offset": 2
            }
        ]
    }'

    If the request is successful, the Kafka Bridge returns a 204 code only.

  2. Submit a GET request to the records endpoint:

    curl -X GET http://localhost:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer/records \
      -H 'accept: application/vnd.kafka.json.v2+json'

    The Kafka Bridge returns messages from the offset that you seeked to.

  3. Restore the default message retrieval behavior by seeking to the last offset for the same partition. This time, use the positions/end endpoint.

    curl -X POST http://localhost:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer/positions/end \
      -H 'content-type: application/vnd.kafka.v2+json' \
      -d '{
        "partitions": [
            {
                "topic": "bridge-quickstart-topic",
                "partition": 0
            }
        ]
    }'

    If the request is successful, the Kafka Bridge returns another 204 code.

Note
You can also use the positions/beginning endpoint to seek to the first offset for one or more partitions.
What to do next

In this quickstart, you have used the Strimzi Kafka Bridge to perform several common operations on a Kafka cluster. You can now delete the Kafka Bridge consumer that you created earlier.

Additional resources

6.2.9. Deleting a Kafka Bridge consumer

Finally, delete the Kafa Bridge consumer that you used throughout this quickstart.

Procedure
  • Delete the Kafka Bridge consumer by sending a DELETE request to the instances endpoint.

    curl -X DELETE http://localhost:8080/consumers/bridge-quickstart-consumer-group/instances/bridge-quickstart-consumer

    If the request is successful, the Kafka Bridge returns a 204 code only.

Additional resources

7. Cruise Control for cluster rebalancing

You can deploy Cruise Control to your Strimzi cluster and use it to rebalance the Kafka cluster.

Cruise Control is an open source system for automating Kafka operations, such as monitoring cluster workload, rebalancing a cluster based on predefined constraints, and detecting and fixing anomalies. It consists of four main components—​the Load Monitor, the Analyzer, the Anomaly Detector, and the Executor—​and a REST API for client interactions. Strimzi utilizes the REST API to support the following Cruise Control features:

  • Generating optimization proposals from multiple optimization goals.

  • Rebalancing a Kafka cluster based on an optimization proposal.

Other Cruise Control features are not currently supported, including anomaly detection, notifications, write-your-own goals, and changing the topic replication factor.

Example YAML files for Cruise Control are provided in examples/cruise-control/.

7.1. Why use Cruise Control?

Cruise Control reduces the time and effort involved in running an efficient and balanced Kafka cluster.

A typical cluster can become unevenly loaded over time. Partitions that handle large amounts of message traffic might be unevenly distributed across the available brokers. To rebalance the cluster, administrators must monitor the load on brokers and manually reassign busy partitions to brokers with spare capacity.

Cruise Control automates the cluster rebalancing process. It constructs a workload model of resource utilization for the cluster—​based on CPU, disk, and network load—​and generates optimization proposals (that you can approve or reject) for more balanced partition assignments. A set of configurable optimization goals is used to calculate these proposals.

When you approve an optimization proposal, Cruise Control applies it to your Kafka cluster. When the cluster rebalancing operation is complete, the broker pods are used more effectively and the Kafka cluster is more evenly balanced.

Additional resources

7.2. Optimization goals overview

To rebalance a Kafka cluster, Cruise Control uses optimization goals to generate optimization proposals, which you can approve or reject.

Optimization goals are constraints on workload redistribution and resource utilization across a Kafka cluster. Strimzi supports most of the optimization goals developed in the Cruise Control project. The supported goals, in the default descending order of priority, are as follows:

  1. Rack-awareness

  2. Minimum number of leader replicas per broker for a set of topics

  3. Replica capacity

  4. Capacity: Disk capacity, Network inbound capacity, Network outbound capacity, CPU capacity

  5. Replica distribution

  6. Potential network output

  7. Resource distribution: Disk utilization distribution, Network inbound utilization distribution, Network outbound utilization distribution, CPU utilization distribution

    Note

    The resource distribution goals are controlled using capacity limits on broker resources.

  8. Leader bytes-in rate distribution

  9. Topic replica distribution

  10. Leader replica distribution

  11. Preferred leader election

For more information on each optimization goal, see Goals in the Cruise Control Wiki.

Note
Intra-broker disk goals, "Write your own" goals, and Kafka assigner goals are not yet supported.

Goals configuration in Strimzi custom resources

You configure optimization goals in Kafka and KafkaRebalance custom resources. Cruise Control has configurations for hard optimization goals that must be satisfied, as well as main, default, and user-provided optimization goals. Optimization goals for resource distribution (disk, network inbound, network outbound, and CPU) are subject to capacity limits on broker resources.

The following sections describe each goal configuration in more detail.

Hard goals and soft goals

Hard goals are goals that must be satisfied in optimization proposals. Goals that are not configured as hard goals are known as soft goals. You can think of soft goals as best effort goals: they do not need to be satisfied in optimization proposals, but are included in optimization calculations. An optimization proposal that violates one or more soft goals, but satisfies all hard goals, is valid.

Cruise Control will calculate optimization proposals that satisfy all the hard goals and as many soft goals as possible (in their priority order). An optimization proposal that does not satisfy all the hard goals is rejected by Cruise Control and not sent to the user for approval.

Note
For example, you might have a soft goal to distribute a topic’s replicas evenly across the cluster (the topic replica distribution goal). Cruise Control will ignore this goal if doing so enables all the configured hard goals to be met.

In Cruise Control, the following main optimization goals are preset as hard goals:

RackAwareGoal; MinTopicLeadersPerBrokerGoal; ReplicaCapacityGoal; DiskCapacityGoal; NetworkInboundCapacityGoal; NetworkOutboundCapacityGoal; CpuCapacityGoal

You configure hard goals in the Cruise Control deployment configuration, by editing the hard.goals property in Kafka.spec.cruiseControl.config.

  • To inherit the preset hard goals from Cruise Control, do not specify the hard.goals property in Kafka.spec.cruiseControl.config

  • To change the preset hard goals, specify the desired goals in the hard.goals property, using their fully-qualified domain names.

Example Kafka configuration for hard optimization goals
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
  zookeeper:
    # ...
  entityOperator:
    topicOperator: {}
    userOperator: {}
  cruiseControl:
    brokerCapacity:
      inboundNetwork: 10000KB/s
      outboundNetwork: 10000KB/s
    config:
      hard.goals: >
        com.linkedin.kafka.cruisecontrol.analyzer.goals.NetworkInboundCapacityGoal,
        com.linkedin.kafka.cruisecontrol.analyzer.goals.NetworkOutboundCapacityGoal
      # ...

Increasing the number of configured hard goals will reduce the likelihood of Cruise Control generating valid optimization proposals.

If skipHardGoalCheck: true is specified in the KafkaRebalance custom resource, Cruise Control does not check that the list of user-provided optimization goals (in KafkaRebalance.spec.goals) contains all the configured hard goals (hard.goals). Therefore, if some, but not all, of the user-provided optimization goals are in the hard.goals list, Cruise Control will still treat them as hard goals even if skipHardGoalCheck: true is specified.

Main optimization goals

The main optimization goals are available to all users. Goals that are not listed in the main optimization goals are not available for use in Cruise Control operations.

Unless you change the Cruise Control deployment configuration, Strimzi will inherit the following main optimization goals from Cruise Control, in descending priority order:

RackAwareGoal; ReplicaCapacityGoal; DiskCapacityGoal; NetworkInboundCapacityGoal; NetworkOutboundCapacityGoal; CpuCapacityGoal; ReplicaDistributionGoal; PotentialNwOutGoal; DiskUsageDistributionGoal; NetworkInboundUsageDistributionGoal; NetworkOutboundUsageDistributionGoal; CpuUsageDistributionGoal; TopicReplicaDistributionGoal; LeaderReplicaDistributionGoal; LeaderBytesInDistributionGoal; PreferredLeaderElectionGoal

Six of these goals are preset as hard goals.

To reduce complexity, we recommend that you use the inherited main optimization goals, unless you need to completely exclude one or more goals from use in KafkaRebalance resources. The priority order of the main optimization goals can be modified, if desired, in the configuration for default optimization goals.

You configure main optimization goals, if necessary, in the Cruise Control deployment configuration: Kafka.spec.cruiseControl.config.goals

  • To accept the inherited main optimization goals, do not specify the goals property in Kafka.spec.cruiseControl.config.

  • If you need to modify the inherited main optimization goals, specify a list of goals, in descending priority order, in the goals configuration option.

Note
If you change the inherited main optimization goals, you must ensure that the hard goals, if configured in the hard.goals property in Kafka.spec.cruiseControl.config, are a subset of the main optimization goals that you configured. Otherwise, errors will occur when generating optimization proposals.
Default optimization goals

Cruise Control uses the default optimization goals to generate the cached optimization proposal. For more information about the cached optimization proposal, see Optimization proposals overview.

You can override the default optimization goals by setting user-provided optimization goals in a KafkaRebalance custom resource.

Unless you specify default.goals in the Cruise Control deployment configuration, the main optimization goals are used as the default optimization goals. In this case, the cached optimization proposal is generated using the main optimization goals.

  • To use the main optimization goals as the default goals, do not specify the default.goals property in Kafka.spec.cruiseControl.config.

  • To modify the default optimization goals, edit the default.goals property in Kafka.spec.cruiseControl.config. You must use a subset of the main optimization goals.

Example Kafka configuration for default optimization goals
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
  zookeeper:
    # ...
  entityOperator:
    topicOperator: {}
    userOperator: {}
  cruiseControl:
    brokerCapacity:
      inboundNetwork: 10000KB/s
      outboundNetwork: 10000KB/s
    config:
      default.goals: >
         com.linkedin.kafka.cruisecontrol.analyzer.goals.RackAwareGoal,
         com.linkedin.kafka.cruisecontrol.analyzer.goals.ReplicaCapacityGoal,
         com.linkedin.kafka.cruisecontrol.analyzer.goals.DiskCapacityGoal
      # ...

If no default optimization goals are specified, the cached proposal is generated using the main optimization goals.

User-provided optimization goals

User-provided optimization goals narrow down the configured default goals for a particular optimization proposal. You can set them, as required, in spec.goals in a KafkaRebalance custom resource:

KafkaRebalance.spec.goals

User-provided optimization goals can generate optimization proposals for different scenarios. For example, you might want to optimize leader replica distribution across the Kafka cluster without considering disk capacity or disk utilization. So, you create a KafkaRebalance custom resource containing a single user-provided goal for leader replica distribution.

User-provided optimization goals must:

  • Include all configured hard goals, or an error occurs

  • Be a subset of the main optimization goals

To ignore the configured hard goals when generating an optimization proposal, add the skipHardGoalCheck: true property to the KafkaRebalance custom resource. See Generating optimization proposals.

Additional resources

7.3. Optimization proposals overview

An optimization proposal is a summary of proposed changes that would produce a more balanced Kafka cluster, with partition workloads distributed more evenly among the brokers. Each optimization proposal is based on the set of optimization goals that was used to generate it, subject to any configured capacity limits on broker resources.

An optimization proposal is contained in the Status.Optimization Result property of a KafkaRebalance custom resource. The information provided is a summary of the full optimization proposal. Use the summary to decide whether to:

  • Approve the optimization proposal. This instructs Cruise Control to apply the proposal to the Kafka cluster and start a cluster rebalance operation.

  • Reject the optimization proposal. You can change the optimization goals and then generate another proposal.

All optimization proposals are dry runs: you cannot approve a cluster rebalance without first generating an optimization proposal. There is no limit to the number of optimization proposals that can be generated.

Cached optimization proposal

Cruise Control maintains a cached optimization proposal based on the configured default optimization goals. Generated from the workload model, the cached optimization proposal is updated every 15 minutes to reflect the current state of the Kafka cluster. If you generate an optimization proposal using the default optimization goals, Cruise Control returns the most recent cached proposal.

To change the cached optimization proposal refresh interval, edit the proposal.expiration.ms setting in the Cruise Control deployment configuration. Consider a shorter interval for fast changing clusters, although this increases the load on the Cruise Control server.

Contents of optimization proposals

An optimization proposal consists of two main sections:

  • The summary is stored in the status of the KafkaRebalance resource.

  • The broker load is stored in a ConfigMap that contains data as a JSON string.

The summary provides an overview of the proposed cluster rebalance and indicates the scale of the changes involved. The broker load shows before and after values for the proposed rebalance, so you can see the impact on each of the brokers in the cluster.

Summary

The following table explains the properties contained in the optimization proposal’s summary section:

Table 2. Properties contained in an optimization proposal
JSON property Description

numIntraBrokerReplicaMovements

The total number of partition replicas that will be transferred between the disks of the cluster’s brokers.

Performance impact during rebalance operation: Relatively high, but lower than numReplicaMovements.

excludedBrokersForLeadership

Not yet supported. An empty list is returned.

numReplicaMovements

The number of partition replicas that will be moved between separate brokers.

Performance impact during rebalance operation: Relatively high.

onDemandBalancednessScoreBefore, onDemandBalancednessScoreAfter

A measurement of the overall balancedness of a Kafka Cluster, before and after the optimization proposal was generated.

The score is calculated by subtracting the sum of the BalancednessScore of each violated soft goal from 100. Cruise Control assigns a BalancednessScore to every optimization goal based on several factors, including priority—​the goal’s position in the list of default.goals or user-provided goals.

The Before score is based on the current configuration of the Kafka cluster. The After score is based on the generated optimization proposal.

intraBrokerDataToMoveMB

The sum of the size of each partition replica that will be moved between disks on the same broker (see also numIntraBrokerReplicaMovements).

Performance impact during rebalance operation: Variable. The larger the number, the longer the cluster rebalance will take to complete. Moving a large amount of data between disks on the same broker has less impact than between separate brokers (see dataToMoveMB).

recentWindows

The number of metrics windows upon which the optimization proposal is based.

dataToMoveMB

The sum of the size of each partition replica that will be moved to a separate broker (see also numReplicaMovements).

Performance impact during rebalance operation: Variable. The larger the number, the longer the cluster rebalance will take to complete.

monitoredPartitionsPercentage

The percentage of partitions in the Kafka cluster covered by the optimization proposal. Affected by the number of excludedTopics.

excludedTopics

If you specified a regular expression in the spec.excludedTopicsRegex property in the KafkaRebalance resource, all topic names matching that expression are listed here. These topics are excluded from the calculation of partition replica/leader movements in the optimization proposal.

numLeaderMovements

The number of partitions whose leaders will be switched to different replicas. This involves a change to ZooKeeper configuration.

Performance impact during rebalance operation: Relatively low.

excludedBrokersForReplicaMove

Not yet supported. An empty list is returned.

Broker load

The broker load is stored in a ConfigMap (with the same name as the KafkaRebalance custom resource) as a JSON formatted string. This JSON string consists of a JSON object with keys for each broker IDs linking to a number of metrics for each broker. Each metric consist of three values. The first is the metric value before the optimization proposal is applied, the second is the expected value of the metric after the proposal is applied, and the third is the difference between the first two values (after minus before).

To extract the JSON string from the ConfigMap you could use the following command, which uses the jq command line JSON parser tool:

kubectl get configmap MY-REBALANCE -o json | jq '.["data"]["brokerLoad.json"]|fromjson|.'

The following table explains the properties contained in the optimization proposal’s broker load ConfigMap:

JSON property Description

leaders

The number of replicas on this broker that are partition leaders.

replicas

The number of replicas on this broker.

cpuPercentage

The CPU utilization as a percentage of the defined capacity.

diskUsedPercentage

The disk utilization as a percentage of the defined capacity.

diskUsedMB

The absolute disk usage in MB.

networkOutRate

The total network output rate for the broker.

leaderNetworkInRate

The network input rate for all partition leader replicas on this broker.

followerNetworkInRate

The network input rate for all follower replicas on this broker.

potentialMaxNetworkOutRate

The hypothetical maximum network output rate that would be realized if this broker became the leader of all the replicas it currently hosts.

7.4. Rebalance performance tuning overview

You can adjust several performance tuning options for cluster rebalances. These options control how partition replica and leadership movements in a rebalance are executed, as well as the bandwidth that is allocated to a rebalance operation.

Partition reassignment commands

Optimization proposals are comprised of separate partition reassignment commands. When you approve a proposal, the Cruise Control server applies these commands to the Kafka cluster.

A partition reassignment command consists of either of the following types of operations:

  • Partition movement: Involves transferring the partition replica and its data to a new location. Partition movements can take one of two forms:

    • Inter-broker movement: The partition replica is moved to a log directory on a different broker.

    • Intra-broker movement: The partition replica is moved to a different log directory on the same broker.

  • Leadership movement: This involves switching the leader of the partition’s replicas.

Cruise Control issues partition reassignment commands to the Kafka cluster in batches. The performance of the cluster during the rebalance is affected by the number of each type of movement contained in each batch.

Replica movement strategies

Cluster rebalance performance is also influenced by the replica movement strategy that is applied to the batches of partition reassignment commands. By default, Cruise Control uses the BaseReplicaMovementStrategy, which simply applies the commands in the order they were generated. However, if there are some very large partition reassignments early in the proposal, this strategy can slow down the application of the other reassignments.

Cruise Control provides four alternative replica movement strategies that can be applied to optimization proposals:

  • PrioritizeSmallReplicaMovementStrategy: Order reassignments in order of ascending size.

  • PrioritizeLargeReplicaMovementStrategy: Order reassignments in order of descending size.

  • PostponeUrpReplicaMovementStrategy: Prioritize reassignments for replicas of partitions which have no out-of-sync replicas.

  • PrioritizeMinIsrWithOfflineReplicasStrategy: Prioritize reassignments with (At/Under)MinISR partitions with offline replicas. This strategy will only work if cruiseControl.config.concurrency.adjuster.min.isr.check.enabled is set to true in the Kafka custom resource’s spec.

These strategies can be configured as a sequence. The first strategy attempts to compare two partition reassignments using its internal logic. If the reassignments are equivalent, then it passes them to the next strategy in the sequence to decide the order, and so on.

Rebalance tuning options

Cruise Control provides several configuration options for tuning the rebalance parameters discussed above. You can set these tuning options at either the Cruise Control server or optimization proposal levels:

  • The Cruise Control server setting can be set in the Kafka custom resource under Kafka.spec.cruiseControl.config.

  • The individual rebalance performance configurations can be set under KafkaRebalance.spec.

The relevant configurations are summarized below:

Server and KafkaRebalance Configuration Description Default Value

num.concurrent.partition.movements.per.broker

The maximum number of inter-broker partition movements in each partition reassignment batch

5

concurrentPartitionMovementsPerBroker

num.concurrent.intra.broker.partition.movements

The maximum number of intra-broker partition movements in each partition reassignment batch

2

concurrentIntraBrokerPartitionMovements

num.concurrent.leader.movements

The maximum number of partition leadership changes in each partition reassignment batch

1000

concurrentLeaderMovements

default.replication.throttle

The bandwidth (in bytes per second) to be assigned to the reassigning of partitions

No Limit

replicationThrottle

default.replica.movement.strategies

The list of strategies (in priority order) used to determine the order in which partition reassignment commands are executed for generated proposals.

For the server setting, use a comma separated string with the fully qualified names of the strategy class (add com.linkedin.kafka.cruisecontrol.executor.strategy. to the start of each class name). For the KafkaRebalance resource setting use a YAML array of strategy class names.

BaseReplicaMovementStrategy

replicaMovementStrategies

Changing the default settings affects the length of time that the rebalance takes to complete, as well as the load placed on the Kafka cluster during the rebalance. Using lower values reduces the load but increases the amount of time taken, and vice versa.

7.5. Cruise Control configuration

The config property in Kafka.spec.cruiseControl contains configuration options as keys with values as one of the following JSON types:

  • String

  • Number

  • Boolean

You can specify and configure all the options listed in the "Configurations" section of the Cruise Control documentation, apart from those managed directly by Strimzi. Specifically, you cannot modify configuration options with keys equal to or starting with one of the keys mentioned here.

If restricted options are specified, they are ignored and a warning message is printed to the Cluster Operator log file. All the supported options are passed to Cruise Control.

An example Cruise Control configuration
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  # ...
  cruiseControl:
    # ...
    config:
      default.goals: >
         com.linkedin.kafka.cruisecontrol.analyzer.goals.RackAwareGoal,
         com.linkedin.kafka.cruisecontrol.analyzer.goals.ReplicaCapacityGoal
      cpu.balance.threshold: 1.1
      metadata.max.age.ms: 300000
      send.buffer.bytes: 131072
    # ...

Cross-Origin Resource Sharing configuration

Cross-Origin Resource Sharing (CORS) allows you to specify allowed methods and originating URLs for accessing REST APIs.

By default, CORS is disabled for the Cruise Control REST API. When enabled, only GET requests for read-only access to the Kafka cluster state are allowed. This means that external applications, which are running in different origins than the Strimzi components, cannot make POST requests to the Cruise Control API. However, those applications can make GET requests to access read-only information about the Kafka cluster, such as the current cluster load or the most recent optimization proposal.

Enabling CORS for Cruise Control

You enable and configure CORS in Kafka.spec.cruiseControl.config.

apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  # ...
  cruiseControl:
    # ...
    config:
      webserver.http.cors.enabled: true
      webserver.http.cors.origin: "*"
      webserver.http.cors.exposeheaders: "User-Task-ID,Content-Type"
    # ...

For more information, see REST APIs in the Cruise Control Wiki.

Capacity configuration

Cruise Control uses capacity limits to determine if optimization goals for resource distribution are being broken. There are four goals of this type:

  • DiskUsageDistributionGoal - Disk utilization distribution

  • CpuUsageDistributionGoal - CPU utilization distribution

  • NetworkInboundUsageDistributionGoal - Network inbound utilization distribution

  • NetworkOutboundUsageDistributionGoal - Network outbound utilization distribution

You specify capacity limits for Kafka broker resources in the brokerCapacity property in Kafka.spec.cruiseControl . They are enabled by default and you can change their default values. Capacity limits can be set for the following broker resources, using the standard Kubernetes byte units (K, M, G and T) or their bibyte (power of two) equivalents (Ki, Mi, Gi and Ti):

  • disk - Disk storage per broker (Default: 100000Mi)

  • cpuUtilization - CPU utilization as a percentage (Default: 100)

  • inboundNetwork - Inbound network throughput in byte units per second (Default: 10000KiB/s)

  • outboundNetwork - Outbound network throughput in byte units per second (Default: 10000KiB/s)

Because Strimzi Kafka brokers are homogeneous, Cruise Control applies the same capacity limits to every broker it is monitoring.

An example Cruise Control brokerCapacity configuration using bibyte units
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  # ...
  cruiseControl:
    # ...
    brokerCapacity:
      disk: 100Gi
      cpuUtilization: 100
      inboundNetwork: 10000KiB/s
      outboundNetwork: 10000KiB/s
    # ...
Additional resources

For more information, refer to the BrokerCapacity schema reference.

Logging configuration

Cruise Control has its own configurable logger:

  • rootLogger.level

Cruise Control uses the Apache log4j 2 logger implementation.

Use the logging property to configure loggers and logger levels.

You can set the log levels by specifying the logger and level directly (inline) or use a custom (external) ConfigMap. If a ConfigMap is used, you set logging.valueFrom.configMapKeyRef.name property to the name of the ConfigMap containing the external logging configuration. Inside the ConfigMap, the logging configuration is described using log4j.properties. Both logging.valueFrom.configMapKeyRef.name and logging.valueFrom.configMapKeyRef.key properties are mandatory. A ConfigMap using the exact logging configuration specified is created with the custom resource when the Cluster Operator is running, then recreated after each reconciliation. If you do not specify a custom ConfigMap, default logging settings are used. If a specific logger value is not set, upper-level logger settings are inherited for that logger. Here we see examples of inline and external logging.

Inline logging
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
# ...
spec:
  cruiseControl:
    # ...
    logging:
      type: inline
      loggers:
        rootLogger.level: "INFO"
    # ...
External logging
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
# ...
spec:
  cruiseControl:
    # ...
    logging:
      type: external
      valueFrom:
        configMapKeyRef:
          name: customConfigMap
          key: cruise-control-log4j.properties
    # ...

7.6. Deploying Cruise Control

To deploy Cruise Control to your Strimzi cluster, define the configuration using the cruiseControl property in the Kafka resource, and then create or update the resource.

Deploy one instance of Cruise Control per Kafka cluster.

Prerequisites
  • A Kubernetes cluster

  • A running Cluster Operator

Procedure
  1. Edit the Kafka resource and add the cruiseControl property.

    The properties you can configure are shown in this example configuration:

    apiVersion: kafka.strimzi.io/v1beta2
    kind: Kafka
    metadata:
      name: my-cluster
    spec:
      # ...
      cruiseControl:
        brokerCapacity: (1)
          inboundNetwork: 10000KB/s
          outboundNetwork: 10000KB/s
          # ...
        config: (2)
          default.goals: >
             com.linkedin.kafka.cruisecontrol.analyzer.goals.RackAwareGoal,
             com.linkedin.kafka.cruisecontrol.analyzer.goals.ReplicaCapacityGoal
             # ...
          cpu.balance.threshold: 1.1
          metadata.max.age.ms: 300000
          send.buffer.bytes: 131072
          # ...
        resources: (3)
          requests:
            cpu: 1
            memory: 512Mi
          limits:
            cpu: 2
            memory: 2Gi
        logging: (4)
            type: inline
            loggers:
              rootLogger.level: "INFO"
        template: (5)
          pod:
            metadata:
              labels:
                label1: value1
            securityContext:
              runAsUser: 1000001
              fsGroup: 0
            terminationGracePeriodSeconds: 120
        readinessProbe: (6)
          initialDelaySeconds: 15
          timeoutSeconds: 5
        livenessProbe: (7)
          initialDelaySeconds: 15
          timeoutSeconds: 5
    # ...
    1. Specifies capacity limits for broker resources. For more information, see Capacity configuration.

    2. Defines the Cruise Control configuration, including the default optimization goals (in default.goals) and any customizations to the main optimization goals (in goals) or the hard goals (in hard.goals). You can provide any standard Cruise Control configuration option apart from those managed directly by Strimzi. For more information on configuring optimization goals, see Optimization goals overview.

    3. CPU and memory resources reserved for Cruise Control. For more information, see resources.

    4. Defined loggers and log levels added directly (inline) or indirectly (external) through a ConfigMap. A custom ConfigMap must be placed under the log4j.properties key. Cruise Control has a single logger named rootLogger.level. You can set the log level to INFO, ERROR, WARN, TRACE, DEBUG, FATAL or OFF. For more information, see Logging configuration.

    5. Customization of deployment templates and pods.

    6. Healthcheck readiness probes.

    7. Healthcheck liveness probes.

  2. Create or update the resource:

    kubectl apply -f kafka.yaml
  3. Verify that Cruise Control was successfully deployed:

    kubectl get deployments -l app.kubernetes.io/name=cruise-control

Auto-created topics

The following table shows the three topics that are automatically created when Cruise Control is deployed. These topics are required for Cruise Control to work properly and must not be deleted or changed.

Table 3. Auto-created topics
Auto-created topic Created by Function

strimzi.cruisecontrol.metrics

Strimzi Metrics Reporter

Stores the raw metrics from the Metrics Reporter in each Kafka broker.

strimzi.cruisecontrol.partitionmetricsamples

Cruise Control

Stores the derived metrics for each partition. These are created by the Metric Sample Aggregator.

strimzi.cruisecontrol.modeltrainingsamples

Cruise Control

Stores the metrics samples used to create the Cluster Workload Model.

To prevent the removal of records that are needed by Cruise Control, log compaction is disabled in the auto-created topics.

What to do next

After configuring and deploying Cruise Control, you can generate optimization proposals.

7.7. Generating optimization proposals

When you create or update a KafkaRebalance resource, Cruise Control generates an optimization proposal for the Kafka cluster based on the configured optimization goals.

Analyze the information in the optimization proposal and decide whether to approve it.

Prerequisites
Procedure
  1. Create a KafkaRebalance resource:

    1. To use the default optimization goals defined in the Kafka resource, leave the spec property empty:

      apiVersion: kafka.strimzi.io/v1beta2
      kind: KafkaRebalance
      metadata:
        name: my-rebalance
        labels:
          strimzi.io/cluster: my-cluster
      spec: {}
    2. To configure user-provided optimization goals instead of using the default goals, add the goals property and enter one or more goals.

      In the following example, rack awareness and replica capacity are configured as user-provided optimization goals:

      apiVersion: kafka.strimzi.io/v1beta2
      kind: KafkaRebalance
      metadata:
        name: my-rebalance
        labels:
          strimzi.io/cluster: my-cluster
      spec:
        goals:
          - RackAwareGoal
          - ReplicaCapacityGoal
    3. To ignore the configured hard goals, add the skipHardGoalCheck: true property:

      apiVersion: kafka.strimzi.io/v1beta2
      kind: KafkaRebalance
      metadata:
        name: my-rebalance
        labels:
          strimzi.io/cluster: my-cluster
      spec:
        goals:
          - RackAwareGoal
          - ReplicaCapacityGoal
        skipHardGoalCheck: true
  2. Create or update the resource:

    kubectl apply -f your-file

    The Cluster Operator requests the optimization proposal from Cruise Control. This might take a few minutes depending on the size of the Kafka cluster.

  3. Check the status of the KafkaRebalance resource:

    kubectl describe kafkarebalance rebalance-cr-name

    Cruise Control returns one of two statuses:

    • PendingProposal: The rebalance operator is polling the Cruise Control API to check if the optimization proposal is ready.

    • ProposalReady: The optimization proposal is ready for review and, if desired, approval. The optimization proposal is contained in the Status.Optimization Result property of the KafkaRebalance resource.

  4. Review the optimization proposal.

    kubectl describe kafkarebalance rebalance-cr-name

    Here is an example proposal:

    Status:
      Conditions:
        Last Transition Time:  2020-05-19T13:50:12.533Z
        Status:                ProposalReady
        Type:                  State
      Observed Generation:     1
      Optimization Result:
        Data To Move MB:  0
        Excluded Brokers For Leadership:
        Excluded Brokers For Replica Move:
        Excluded Topics:
        Intra Broker Data To Move MB:         0
        Monitored Partitions Percentage:      100
        Num Intra Broker Replica Movements:   0
        Num Leader Movements:                 0
        Num Replica Movements:                26
        On Demand Balancedness Score After:   81.8666802863978
        On Demand Balancedness Score Before:  78.01176356230222
        Recent Windows:                       1
      Session Id:                             05539377-ca7b-45ef-b359-e13564f1458c

    The properties in the Optimization Result section describe the pending cluster rebalance operation. For descriptions of each property, see Contents of optimization proposals.

Additional resources

7.8. Approving an optimization proposal

You can approve an optimization proposal generated by Cruise Control, if its status is ProposalReady. Cruise Control will then apply the optimization proposal to the Kafka cluster, reassigning partitions to brokers and changing partition leadership.

Caution

This is not a dry run. Before you approve an optimization proposal, you must:

Prerequisites
Procedure

Perform these steps for the optimization proposal that you want to approve:

  1. Unless the optimization proposal is newly generated, check that it is based on current information about the state of the Kafka cluster. To do so, refresh the optimization proposal to make sure it uses the latest cluster metrics:

    1. Annotate the KafkaRebalance resource in Kubernetes with refresh:

      kubectl annotate kafkarebalance rebalance-cr-name strimzi.io/rebalance=refresh
    2. Check the status of the KafkaRebalance resource:

      kubectl describe kafkarebalance rebalance-cr-name
    3. Wait until the status changes to ProposalReady.

  2. Approve the optimization proposal that you want Cruise Control to apply.

    Annotate the KafkaRebalance resource in Kubernetes:

    kubectl annotate kafkarebalance rebalance-cr-name strimzi.io/rebalance=approve
  3. The Cluster Operator detects the annotated resource and instructs Cruise Control to rebalance the Kafka cluster.

  4. Check the status of the KafkaRebalance resource:

    kubectl describe kafkarebalance rebalance-cr-name
  5. Cruise Control returns one of three statuses:

    • Rebalancing: The cluster rebalance operation is in progress.

    • Ready: The cluster rebalancing operation completed successfully. To use the same KafkaRebalance custom resource to generate another optimization proposal, apply the refresh annotation to the custom resource. This moves the custom resource to the PendingProposal or ProposalReady state. You can then review the optimization proposal and approve it, if desired.

    • NotReady: An error occurred—​see Fixing problems with a KafkaRebalance resource.

7.9. Stopping a cluster rebalance

Once started, a cluster rebalance operation might take some time to complete and affect the overall performance of the Kafka cluster.

If you want to stop a cluster rebalance operation that is in progress, apply the stop annotation to the KafkaRebalance custom resource. This instructs Cruise Control to finish the current batch of partition reassignments and then stop the rebalance. When the rebalance has stopped, completed partition reassignments have already been applied; therefore, the state of the Kafka cluster is different when compared to prior to the start of the rebalance operation. If further rebalancing is required, you should generate a new optimization proposal.

Note
The performance of the Kafka cluster in the intermediate (stopped) state might be worse than in the initial state.
Prerequisites
  • You have approved the optimization proposal by annotating the KafkaRebalance custom resource with approve.

  • The status of the KafkaRebalance custom resource is Rebalancing.

Procedure
  1. Annotate the KafkaRebalance resource in Kubernetes:

    kubectl annotate kafkarebalance rebalance-cr-name strimzi.io/rebalance=stop
  2. Check the status of the KafkaRebalance resource:

    kubectl describe kafkarebalance rebalance-cr-name
  3. Wait until the status changes to Stopped.

Additional resources

7.10. Fixing problems with a KafkaRebalance resource

If an issue occurs when creating a KafkaRebalance resource or interacting with Cruise Control, the error is reported in the resource status, along with details of how to fix it. The resource also moves to the NotReady state.

To continue with the cluster rebalance operation, you must fix the problem in the KafkaRebalance resource itself or with the overall Cruise Control deployment. Problems might include the following:

  • A misconfigured parameter in the KafkaRebalance resource.

  • The strimzi.io/cluster label for specifying the Kafka cluster in the KafkaRebalance resource is missing.

  • The Cruise Control server is not deployed as the cruiseControl property in the Kafka resource is missing.

  • The Cruise Control server is not reachable.

After fixing the issue, you need to add the refresh annotation to the KafkaRebalance resource. During a “refresh”, a new optimization proposal is requested from the Cruise Control server.

Prerequisites
Procedure
  1. Get information about the error from the KafkaRebalance status:

    kubectl describe kafkarebalance rebalance-cr-name
  2. Attempt to resolve the issue in the KafkaRebalance resource.

  3. Annotate the KafkaRebalance resource in Kubernetes:

    kubectl annotate kafkarebalance rebalance-cr-name strimzi.io/rebalance=refresh
  4. Check the status of the KafkaRebalance resource:

    kubectl describe kafkarebalance rebalance-cr-name
  5. Wait until the status changes to PendingProposal, or directly to ProposalReady.

Additional resources

8. Distributed tracing

Distributed tracing allows you to track the progress of transactions between applications in a distributed system. In a microservices architecture, tracing tracks the progress of transactions between services. Trace data is useful for monitoring application performance and investigating issues with target systems and end-user applications.

In Strimzi, tracing facilitates the end-to-end tracking of messages: from source systems to Kafka, and then from Kafka to target systems and applications. It complements the metrics that are available to view in Grafana dashboards, as well as the component loggers.

How Strimzi supports tracing

Support for tracing is built in to the following components:

  • Kafka Connect (including Kafka Connect with Source2Image support)

  • MirrorMaker

  • MirrorMaker 2.0

  • Strimzi Kafka Bridge

You enable and configure tracing for these components using template configuration properties in their custom resources.

To enable tracing in Kafka producers, consumers, and Kafka Streams API applications, you instrument application code using the OpenTracing Apache Kafka Client Instrumentation library (included with Strimzi). When instrumented, clients generate trace data; for example, when producing messages or writing offsets to the log.

Traces are sampled according to a sampling strategy and then visualized in the Jaeger user interface.

Note

Tracing is not supported for Kafka brokers.

Setting up tracing for applications and systems beyond Strimzi is outside the scope of this chapter. To learn more about this subject, search for "inject and extract" in the OpenTracing documentation.

Outline of procedures

To set up tracing for Strimzi, follow these procedures in order:

Prerequisites

8.1. Overview of OpenTracing and Jaeger

Strimzi uses the OpenTracing and Jaeger projects.

OpenTracing is an API specification that is independent from the tracing or monitoring system.

  • The OpenTracing APIs are used to instrument application code

  • Instrumented applications generate traces for individual transactions across the distributed system

  • Traces are composed of spans that define specific units of work over time

Jaeger is a tracing system for microservices-based distributed systems.

  • Jaeger implements the OpenTracing APIs and provides client libraries for instrumentation

  • The Jaeger user interface allows you to query, filter, and analyze trace data

The Jaeger user interface showing a simple query

Additional resources

8.2. Setting up tracing for Kafka clients

Initialize a Jaeger tracer to instrument your client applications for distributed tracing.

8.2.1. Initializing a Jaeger tracer for Kafka clients

Configure and initialize a Jaeger tracer using a set of tracing environment variables.

Procedure

In each client application:

  1. Add Maven dependencies for Jaeger to the pom.xml file for the client application:

    <dependency>
        <groupId>io.jaegertracing</groupId>
        <artifactId>jaeger-client</artifactId>
        <version>1.3.2</version>
    </dependency>
  2. Define the configuration of the Jaeger tracer using the tracing environment variables.

  3. Create the Jaeger tracer from the environment variables that you defined in step two:

    Tracer tracer = Configuration.fromEnv().getTracer();
    Note
    For alternative ways to initialize a Jaeger tracer, see the Java OpenTracing library documentation.
  4. Register the Jaeger tracer as a global tracer:

    GlobalTracer.register(tracer);

A Jaeger tracer is now initialized for the client application to use.

8.2.2. Environment variables for tracing

Use these environment variables when configuring a Jaeger tracer for Kafka clients.

Note
The tracing environment variables are part of the Jaeger project and are subject to change. For the latest environment variables, see the Jaeger documentation.
Property Required Description

JAEGER_SERVICE_NAME

Yes

The name of the Jaeger tracer service.

JAEGER_AGENT_HOST

No

The hostname for communicating with the jaeger-agent through the User Datagram Protocol (UDP).

JAEGER_AGENT_PORT

No

The port used for communicating with the jaeger-agent through UDP.

JAEGER_ENDPOINT

No

The traces endpoint. Only define this variable if the client application will bypass the jaeger-agent and connect directly to the jaeger-collector.

JAEGER_AUTH_TOKEN

No

The authentication token to send to the endpoint as a bearer token.

JAEGER_USER

No

The username to send to the endpoint if using basic authentication.

JAEGER_PASSWORD

No

The password to send to the endpoint if using basic authentication.

JAEGER_PROPAGATION

No

A comma-separated list of formats to use for propagating the trace context. Defaults to the standard Jaeger format. Valid values are jaeger, b3, and w3c.

JAEGER_REPORTER_LOG_SPANS

No

Indicates whether the reporter should also log the spans.

JAEGER_REPORTER_MAX_QUEUE_SIZE

No

The reporter’s maximum queue size.

JAEGER_REPORTER_FLUSH_INTERVAL

No

The reporter’s flush interval, in ms. Defines how frequently the Jaeger reporter flushes span batches.

JAEGER_SAMPLER_TYPE

No

The sampling strategy to use for client traces:

  • Constant

  • Probabilistic

  • Rate Limiting

  • Remote (the default)

To sample all traces, use the Constant sampling strategy with a parameter of 1.

For more information, see the Jaeger documentation.

JAEGER_SAMPLER_PARAM

No

The sampler parameter (number).

JAEGER_SAMPLER_MANAGER_HOST_PORT

No

The hostname and port to use if a Remote sampling strategy is selected.

JAEGER_TAGS

No

A comma-separated list of tracer-level tags that are added to all reported spans.

The value can also refer to an environment variable using the format ${envVarName:default}. :default is optional and identifies a value to use if the environment variable cannot be found.

8.3. Instrumenting Kafka clients with tracers

Instrument Kafka producer and consumer clients, and Kafka Streams API applications for distributed tracing.

8.3.1. Instrumenting producers and consumers for tracing

Use a Decorator pattern or Interceptors to instrument your Java producer and consumer application code for tracing.

Procedure

In the application code of each producer and consumer application:

  1. Add the Maven dependency for OpenTracing to the producer or consumer’s pom.xml file.

    <dependency>
        <groupId>io.opentracing.contrib</groupId>
        <artifactId>opentracing-kafka-client</artifactId>
        <version>0.1.15</version>
    </dependency>
  2. Instrument your client application code using either a Decorator pattern or Interceptors.

    • To use a Decorator pattern:

      // Create an instance of the KafkaProducer:
      KafkaProducer<Integer, String> producer = new KafkaProducer<>(senderProps);
      
      // Create an instance of the TracingKafkaProducer:
      TracingKafkaProducer<Integer, String> tracingProducer = new TracingKafkaProducer<>(producer,
              tracer);
      
      // Send:
      tracingProducer.send(...);
      
      // Create an instance of the KafkaConsumer:
      KafkaConsumer<Integer, String> consumer = new KafkaConsumer<>(consumerProps);
      
      // Create an instance of the TracingKafkaConsumer:
      TracingKafkaConsumer<Integer, String> tracingConsumer = new TracingKafkaConsumer<>(consumer,
              tracer);
      
      // Subscribe:
      tracingConsumer.subscribe(Collections.singletonList("messages"));
      
      // Get messages:
      ConsumerRecords<Integer, String> records = tracingConsumer.poll(1000);
      
      // Retrieve SpanContext from polled record (consumer side):
      ConsumerRecord<Integer, String> record = ...
      SpanContext spanContext = TracingKafkaUtils.extractSpanContext(record.headers(), tracer);
    • To use Interceptors:

      // Register the tracer with GlobalTracer:
      GlobalTracer.register(tracer);
      
      // Add the TracingProducerInterceptor to the sender properties:
      senderProps.put(ProducerConfig.INTERCEPTOR_CLASSES_CONFIG,
                TracingProducerInterceptor.class.getName());
      
      // Create an instance of the KafkaProducer:
      KafkaProducer<Integer, String> producer = new KafkaProducer<>(senderProps);
      
      // Send:
      producer.send(...);
      
      // Add the TracingConsumerInterceptor to the consumer properties:
      consumerProps.put(ConsumerConfig.INTERCEPTOR_CLASSES_CONFIG,
                TracingConsumerInterceptor.class.getName());
      
      // Create an instance of the KafkaConsumer:
      KafkaConsumer<Integer, String> consumer = new KafkaConsumer<>(consumerProps);
      
      // Subscribe:
      consumer.subscribe(Collections.singletonList("messages"));
      
      // Get messages:
      ConsumerRecords<Integer, String> records = consumer.poll(1000);
      
      // Retrieve the SpanContext from a polled message (consumer side):
      ConsumerRecord<Integer, String> record = ...
      SpanContext spanContext = TracingKafkaUtils.extractSpanContext(record.headers(), tracer);
Custom span names in a Decorator pattern

A span is a logical unit of work in Jaeger, with an operation name, start time, and duration.

To use a Decorator pattern to instrument your producer and consumer applications, define custom span names by passing a BiFunction object as an additional argument when creating the TracingKafkaProducer and TracingKafkaConsumer objects. The OpenTracing Apache Kafka Client Instrumentation library includes several built-in span names.

Example: Using custom span names to instrument client application code in a Decorator pattern
// Create a BiFunction for the KafkaProducer that operates on (String operationName, ProducerRecord consumerRecord) and returns a String to be used as the name:

BiFunction<String, ProducerRecord, String> producerSpanNameProvider =
    (operationName, producerRecord) -> "CUSTOM_PRODUCER_NAME";

// Create an instance of the KafkaProducer:
KafkaProducer<Integer, String> producer = new KafkaProducer<>(senderProps);

// Create an instance of the TracingKafkaProducer
TracingKafkaProducer<Integer, String> tracingProducer = new TracingKafkaProducer<>(producer,
        tracer,
        producerSpanNameProvider);

// Spans created by the tracingProducer will now have "CUSTOM_PRODUCER_NAME" as the span name.

// Create a BiFunction for the KafkaConsumer that operates on (String operationName, ConsumerRecord consumerRecord) and returns a String to be used as the name:

BiFunction<String, ConsumerRecord, String> consumerSpanNameProvider =
    (operationName, consumerRecord) -> operationName.toUpperCase();

// Create an instance of the KafkaConsumer:
KafkaConsumer<Integer, String> consumer = new KafkaConsumer<>(consumerProps);

// Create an instance of the TracingKafkaConsumer, passing in the consumerSpanNameProvider BiFunction:

TracingKafkaConsumer<Integer, String> tracingConsumer = new TracingKafkaConsumer<>(consumer,
        tracer,
        consumerSpanNameProvider);

// Spans created by the tracingConsumer will have the operation name as the span name, in upper-case.
// "receive" -> "RECEIVE"
Built-in span names

When defining custom span names, you can use the following BiFunctions in the ClientSpanNameProvider class. If no spanNameProvider is specified, CONSUMER_OPERATION_NAME and PRODUCER_OPERATION_NAME are used.

BiFunction Description

CONSUMER_OPERATION_NAME, PRODUCER_OPERATION_NAME

Returns the operationName as the span name: "receive" for consumers and "send" for producers.

CONSUMER_PREFIXED_OPERATION_NAME(String prefix), PRODUCER_PREFIXED_OPERATION_NAME(String prefix)

Returns a String concatenation of prefix and operationName.

CONSUMER_TOPIC, PRODUCER_TOPIC

Returns the name of the topic that the message was sent to or retrieved from in the format (record.topic()).

PREFIXED_CONSUMER_TOPIC(String prefix), PREFIXED_PRODUCER_TOPIC(String prefix)

Returns a String concatenation of prefix and the topic name in the format (record.topic()).

CONSUMER_OPERATION_NAME_TOPIC, PRODUCER_OPERATION_NAME_TOPIC

Returns the operation name and the topic name: "operationName - record.topic()".

CONSUMER_PREFIXED_OPERATION_NAME_TOPIC(String prefix), PRODUCER_PREFIXED_OPERATION_NAME_TOPIC(String prefix)

Returns a String concatenation of prefix and "operationName - record.topic()".

8.3.2. Instrumenting Kafka Streams applications for tracing

This section describes how to instrument Kafka Streams API applications for distributed tracing.

Procedure

In each Kafka Streams API application:

  1. Add the opentracing-kafka-streams dependency to the pom.xml file for your Kafka Streams API application:

    <dependency>
        <groupId>io.opentracing.contrib</groupId>
        <artifactId>opentracing-kafka-streams</artifactId>
        <version>0.1.15</version>
    </dependency>
  2. Create an instance of the TracingKafkaClientSupplier supplier interface:

    KafkaClientSupplier supplier = new TracingKafkaClientSupplier(tracer);
  3. Provide the supplier interface to KafkaStreams:

    KafkaStreams streams = new KafkaStreams(builder.build(), new StreamsConfig(config), supplier);
    streams.start();

8.4. Setting up tracing for MirrorMaker, Kafka Connect, and the Kafka Bridge

Distributed tracing is supported for MirrorMaker, MirrorMaker 2.0, Kafka Connect (including Kafka Connect with Source2Image support), and the Strimzi Kafka Bridge.

Tracing in MirrorMaker and MirrorMaker 2.0

For MirrorMaker and MirrorMaker 2.0, messages are traced from the source cluster to the target cluster. The trace data records messages entering and leaving the MirrorMaker or MirrorMaker 2.0 component.

Tracing in Kafka Connect

Only messages produced and consumed by Kafka Connect itself are traced. To trace messages sent between Kafka Connect and external systems, you must configure tracing in the connectors for those systems. For more information, see Configuring Kafka Connect.

Tracing in the Kafka Bridge

Messages produced and consumed by the Kafka Bridge are traced. Incoming HTTP requests from client applications to send and receive messages through the Kafka Bridge are also traced. To have end-to-end tracing, you must configure tracing in your HTTP clients.

8.4.1. Enabling tracing in MirrorMaker, Kafka Connect, and Kafka Bridge resources

Update the configuration of KafkaMirrorMaker, KafkaMirrorMaker2, KafkaConnect, KafkaConnectS2I, and KafkaBridge custom resources to specify and configure a Jaeger tracer service for each resource. Updating a tracing-enabled resource in your Kubernetes cluster triggers two events:

  • Interceptor classes are updated in the integrated consumers and producers in MirrorMaker, MirrorMaker 2.0, Kafka Connect, or the Strimzi Kafka Bridge.

  • For MirrorMaker, MirrorMaker 2.0, and Kafka Connect, the tracing agent initializes a Jaeger tracer based on the tracing configuration defined in the resource.

  • For the Kafka Bridge, a Jaeger tracer based on the tracing configuration defined in the resource is initialized by the Kafka Bridge itself.

Procedure

Perform these steps for each KafkaMirrorMaker, KafkaMirrorMaker2, KafkaConnect, KafkaConnectS2I, and KafkaBridge resource.

  1. In the spec.template property, configure the Jaeger tracer service. For example:

    Jaeger tracer configuration for Kafka Connect
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaConnect
    metadata:
      name: my-connect-cluster
    spec:
      #...
      template:
        connectContainer: (1)
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
      tracing: (2)
        type: jaeger
      #...
    Jaeger tracer configuration for MirrorMaker
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaMirrorMaker
    metadata:
      name: my-mirror-maker
    spec:
      #...
      template:
        mirrorMakerContainer:
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
      tracing:
        type: jaeger
    #...
    Jaeger tracer configuration for MirrorMaker 2.0
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaMirrorMaker2
    metadata:
      name: my-mm2-cluster
    spec:
      #...
      template:
        connectContainer:
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
      tracing:
        type: jaeger
    #...
    Jaeger tracer configuration for the Kafka Bridge
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaBridge
    metadata:
      name: my-bridge
    spec:
      #...
      template:
        bridgeContainer:
          env:
            - name: JAEGER_SERVICE_NAME
              value: my-jaeger-service
            - name: JAEGER_AGENT_HOST
              value: jaeger-agent-name
            - name: JAEGER_AGENT_PORT
              value: "6831"
      tracing:
        type: jaeger
    #...
    1. Use the tracing environment variables as template configuration properties.

    2. Set the spec.tracing.type property to jaeger.

  2. Create or update the resource:

    kubectl apply -f your-file

9. Managing TLS certificates

Strimzi supports encrypted communication between the Kafka and Strimzi components using the TLS protocol. Communication between Kafka brokers (interbroker communication), between ZooKeeper nodes (internodal communication), and between those components and the Strimzi operators is always encrypted. Communication between Kafka clients and Kafka brokers is encrypted according to how the cluster is configured. For the Kafka and Strimzi components, TLS certificates are also used for authentication.

The Cluster Operator automatically sets up and renews TLS certificates to enable encryption and authentication within your cluster. It also sets up other TLS certificates if you want to enable encryption or TLS authentication between Kafka brokers and clients. Certificates provided by users are not renewed.

You can provide your own server certificates, called Kafka listener certificates, for TLS listeners or external listeners which have TLS encryption enabled. For more information, see Kafka listener certificates.

Secure Communication
Figure 5. Example architecture of the communication secured by TLS

9.1. Certificate Authorities

To support encryption, each Strimzi component needs its own private keys and public key certificates. All component certificates are signed by an internal Certificate Authority (CA) called the cluster CA.

Similarly, each Kafka client application connecting to Strimzi using TLS client authentication needs to provide private keys and certificates. A second internal CA, named the clients CA, is used to sign certificates for the Kafka clients.

9.1.1. CA certificates

Both the cluster CA and clients CA have a self-signed public key certificate.

Kafka brokers are configured to trust certificates signed by either the cluster CA or clients CA. Components that clients do not need to connect to, such as ZooKeeper, only trust certificates signed by the cluster CA. Unless TLS encryption for external listeners is disabled, client applications must trust certificates signed by the cluster CA. This is also true for client applications that perform mutual TLS authentication.

By default, Strimzi automatically generates and renews CA certificates issued by the cluster CA or clients CA. You can configure the management of these CA certificates in the Kafka.spec.clusterCa and Kafka.spec.clientsCa objects. Certificates provided by users are not renewed.

You can provide your own CA certificates for the cluster CA or clients CA. For more information, see Installing your own CA certificates. If you provide your own certificates, you must manually renew them when needed.

9.1.2. Installing your own CA certificates

This procedure describes how to install your own CA certificates and keys instead of using the CA certificates and private keys generated by the Cluster Operator.

You can use this procedure to install your own cluster or client CA certificates.

The procedure describes renewal of CA certificates in PEM format. You can also use certificates in PKCS #12 format.

Prerequisites
  • The Cluster Operator is running.

  • A Kafka cluster is not yet deployed.

  • Your own X.509 certificates and keys in PEM format for the cluster CA or clients CA.

    • If you want to use a cluster or clients CA which is not a Root CA, you have to include the whole chain in the certificate file. The chain should be in the following order:

      1. The cluster or clients CA

      2. One or more intermediate CAs

      3. The root CA

    • All CAs in the chain should be configured as a CA in the X509v3 Basic Constraints.

Procedure
  1. Put your CA certificate in the corresponding Secret.

    1. Delete the existing secret:

      kubectl delete secret CA-CERTIFICATE-SECRET

      CA-CERTIFICATE-SECRET is the name of the Secret, which is CLUSTER-NAME-cluster-ca-cert for the cluster CA certificate and CLUSTER-NAME-clients-ca-cert for the clients CA certificate.

      Ignore any "Not Exists" errors.

    2. Create and label the new secret

      kubectl create secret generic CA-CERTIFICATE-SECRET --from-file=ca.crt=CA-CERTIFICATE-FILENAME
  2. Put your CA key in the corresponding Secret.

    1. Delete the existing secret:

      kubectl delete secret CA-KEY-SECRET

      CA-KEY-SECRET is the name of CA key, which is CLUSTER-NAME-cluster-ca for the cluster CA key and CLUSTER-NAME-clients-ca for the clients CA key.

    2. Create the new secret:

      kubectl create secret generic CA-KEY-SECRET --from-file=ca.key=CA-KEY-SECRET-FILENAME
  3. Label the secrets with the labels strimzi.io/kind=Kafka and strimzi.io/cluster=CLUSTER-NAME:

    kubectl label secret CA-CERTIFICATE-SECRET strimzi.io/kind=Kafka strimzi.io/cluster=CLUSTER-NAME
    kubectl label secret CA-KEY-SECRET strimzi.io/kind=Kafka strimzi.io/cluster=CLUSTER-NAME
  4. Create the Kafka resource for your cluster, configuring either the Kafka.spec.clusterCa or the Kafka.spec.clientsCa object to not use generated CAs:

    Example fragment Kafka resource configuring the cluster CA to use certificates you supply for yourself
    kind: Kafka
    version: kafka.strimzi.io/v1beta2
    spec:
      # ...
      clusterCa:
        generateCertificateAuthority: false
Additional resources

9.2. Secrets

Strimzi uses Secrets to store private keys and certificates for Kafka cluster components and clients. Secrets are used for establishing TLS encrypted connections between Kafka brokers, and between brokers and clients. They are also used for mutual TLS authentication.

  • A Cluster Secret contains a cluster CA certificate to sign Kafka broker certificates, and is used by a connecting client to establish a TLS encrypted connection with the Kafka cluster to validate broker identity.

  • A Client Secret contains a client CA certificate for a user to sign its own client certificate to allow mutual authentication against the Kafka cluster. The broker validates the client identity through the client CA certificate itself.

  • A User Secret contains a private key and certificate, which are generated and signed by the client CA certificate when a new user is created. The key and certificate are used for authentication and authorization when accessing the cluster.

Secrets provide private keys and certificates in PEM and PKCS #12 formats. Using private keys and certificates in PEM format means that users have to get them from the Secrets, and generate a corresponding truststore (or keystore) to use in their Java applications. PKCS #12 storage provides a truststore (or keystore) that can be used directly.

All keys are 2048 bits in size.

9.2.1. PKCS #12 storage

PKCS #12 defines an archive file format (.p12) for storing cryptography objects into a single file with password protection. You can use PKCS #12 to manage certificates and keys in one place.

Each Secret contains fields specific to PKCS #12.

  • The .p12 field contains the certificates and keys.

  • The .password field is the password that protects the archive.

9.2.2. Cluster CA Secrets

The following tables describe the Cluster Secrets that are managed by the Cluster Operator in a Kafka cluster.

Only the <cluster>-cluster-ca-cert Secret needs to be used by clients. All other Secrets described only need to be accessed by the Strimzi components. You can enforce this using Kubernetes role-based access controls, if necessary.

Table 4. Fields in the <cluster>-cluster-ca Secret
Field Description

ca.key

The current private key for the cluster CA.

Table 5. Fields in the <cluster>-cluster-ca-cert Secret
Field Description

ca.p12

PKCS #12 archive file for storing certificates and keys.

ca.password

Password for protecting the PKCS #12 archive file.

ca.crt

The current certificate for the cluster CA.

Note
The CA certificates in <cluster>-cluster-ca-cert must be trusted by Kafka client applications so that they validate the Kafka broker certificates when connecting to Kafka brokers over TLS.
Table 6. Fields in the <cluster>-kafka-brokers Secret
Field Description

<cluster>-kafka-<num>.p12

PKCS #12 archive file for storing certificates and keys.

<cluster>-kafka-<num>.password

Password for protecting the PKCS #12 archive file.

<cluster>-kafka-<num>.crt

Certificate for Kafka broker pod <num>. Signed by a current or former cluster CA private key in <cluster>-cluster-ca.

<cluster>-kafka-<num>.key

Private key for Kafka broker pod <num>.

Table 7. Fields in the <cluster>-zookeeper-nodes Secret
Field Description

<cluster>-zookeeper-<num>.p12

PKCS #12 archive file for storing certificates and keys.

<cluster>-zookeeper-<num>.password

Password for protecting the PKCS #12 archive file.

<cluster>-zookeeper-<num>.crt

Certificate for ZooKeeper node <num>. Signed by a current or former cluster CA private key in <cluster>-cluster-ca.

<cluster>-zookeeper-<num>.key

Private key for ZooKeeper pod <num>.

Table 8. Fields in the <cluster>-entity-operator-certs Secret
Field Description

entity-operator_.p12

PKCS #12 archive file for storing certificates and keys.

entity-operator_.password

Password for protecting the PKCS #12 archive file.

entity-operator_.crt

Certificate for TLS communication between the Entity Operator and Kafka or ZooKeeper. Signed by a current or former cluster CA private key in <cluster>-cluster-ca.

entity-operator.key

Private key for TLS communication between the Entity Operator and Kafka or ZooKeeper.

9.2.3. Client CA Secrets

Table 9. Clients CA Secrets managed by the Cluster Operator in <cluster>
Secret name Field within Secret Description

<cluster>-clients-ca

ca.key

The current private key for the clients CA.

<cluster>-clients-ca-cert

ca.p12

PKCS #12 archive file for storing certificates and keys.

ca.password

Password for protecting the PKCS #12 archive file.

ca.crt

The current certificate for the clients CA.

The certificates in <cluster>-clients-ca-cert are those which the Kafka brokers trust.

Note
<cluster>-clients-ca is used to sign certificates of client applications. It needs to be accessible to the Strimzi components and for administrative access if you are intending to issue application certificates without using the User Operator. You can enforce this using Kubernetes role-based access controls if necessary.

9.2.4. Adding labels and annotations to Secrets

By configuring the clusterCaCert template property in the Kafka custom resource, you can add custom labels and annotations to the Cluster CA Secrets created by the Cluster Operator. Labels and annotations are useful for identifying objects and adding contextual information. You configure template properties in Strimzi custom resources.

Example template customization to add labels and annotations to Secrets
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
    template:
      clusterCaCert:
        metadata:
          labels:
            label1: value1
            label2: value2
          annotations:
            annotation1: value1
            annotation2: value2
    # ...

For more information on configuring template properties, see Customizing Kubernetes resources.

9.2.5. Disabling ownerReference in the CA Secrets

By default, the Cluster and Client CA Secrets are created with an ownerReference property that is set to the Kafka custom resource. This means that, when the Kafka custom resource is deleted, the CA secrets are also deleted (garbage collected) by Kubernetes.

If you want to reuse the CA for a new cluster, you can disable the ownerReference by setting the generateSecretOwnerReference property for the Cluster and Client CA Secrets to false in the Kafka configuration. When the ownerReference is disabled, CA Secrets are not deleted by Kubernetes when the corresponding Kafka custom resource is deleted.

Example Kafka configuration with disabled ownerReference for Cluster and Client CAs
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
# ...
spec:
# ...
  clusterCa:
    generateSecretOwnerReference: false
  clientsCa:
    generateSecretOwnerReference: false
# ...

9.2.6. User Secrets

Table 10. Secrets managed by the User Operator
Secret name Field within Secret Description

<user>

user.p12

PKCS #12 archive file for storing certificates and keys.

user.password

Password for protecting the PKCS #12 archive file.

user.crt

Certificate for the user, signed by the clients CA

user.key

Private key for the user

9.3. Certificate renewal and validity periods

Cluster CA and clients CA certificates are only valid for a limited time period, known as the validity period. This is usually defined as a number of days since the certificate was generated.

For CA certificates automatically created by the Cluster Operator, you can configure the validity period of:

  • Cluster CA certificates in Kafka.spec.clusterCa.validityDays

  • Client CA certificates in Kafka.spec.clientsCa.validityDays

The default validity period for both certificates is 365 days. Manually-installed CA certificates should have their own validity periods defined.

When a CA certificate expires, components and clients that still trust that certificate will not accept TLS connections from peers whose certificates were signed by the CA private key. The components and clients need to trust the new CA certificate instead.

To allow the renewal of CA certificates without a loss of service, the Cluster Operator will initiate certificate renewal before the old CA certificates expire.

You can configure the renewal period of the certificates created by the Cluster Operator:

  • Cluster CA certificates in Kafka.spec.clusterCa.renewalDays

  • Client CA certificates in Kafka.spec.clientsCa.renewalDays

The default renewal period for both certificates is 30 days.

The renewal period is measured backwards, from the expiry date of the current certificate.

Validity period against renewal period
Not Before                                     Not After
    |                                              |
    |<--------------- validityDays --------------->|
                              <--- renewalDays --->|

To make a change to the validity and renewal periods after creating the Kafka cluster, you configure and apply the Kafka custom resource, and manually renew the CA certificates. If you do not manually renew the certificates, the new periods will be used the next time the certificate is renewed automatically.

Example Kafka configuration for certificate validity and renewal periods
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
# ...
spec:
# ...
  clusterCa:
    renewalDays: 30
    validityDays: 365
    generateCertificateAuthority: true
  clientsCa:
    renewalDays: 30
    validityDays: 365
    generateCertificateAuthority: true
# ...

The behavior of the Cluster Operator during the renewal period depends on the settings for the certificate generation properties, generateCertificateAuthority and generateCertificateAuthority.

true

If the properties are set to true, a CA certificate is generated automatically by the Cluster Operator, and renewed automatically within the renewal period.

false

If the properties are set to false, a CA certificate is not generated by the Cluster Operator. Use this option if you are installing your own certificates.

9.3.1. Renewal process with automatically generated CA certificates

The Cluster Operator performs the following process to renew CA certificates:

  1. Generate a new CA certificate, but retain the existing key. The new certificate replaces the old one with the name ca.crt within the corresponding Secret.

  2. Generate new client certificates (for ZooKeeper nodes, Kafka brokers, and the Entity Operator). This is not strictly necessary because the signing key has not changed, but it keeps the validity period of the client certificate in sync with the CA certificate.

  3. Restart ZooKeeper nodes so that they will trust the new CA certificate and use the new client certificates.

  4. Restart Kafka brokers so that they will trust the new CA certificate and use the new client certificates.

  5. Restart the Topic and User Operators so that they will trust the new CA certificate and use the new client certificates.

9.3.2. Client certificate renewal

The Cluster Operator is not aware of the client applications using the Kafka cluster.

When connecting to the cluster, and to ensure they operate correctly, client applications must:

  • Trust the cluster CA certificate published in the <cluster>-cluster-ca-cert Secret.

  • Use the credentials published in their <user-name> Secret to connect to the cluster.

    The User Secret provides credentials in PEM and PKCS #12 format, or it can provide a password when using SCRAM-SHA authentication. The User Operator creates the user credentials when a user is created.

You must ensure clients continue to work after certificate renewal. The renewal process depends on how the clients are configured.

If you are provisioning client certificates and keys manually, you must generate new client certificates and ensure the new certificates are used by clients within the renewal period. Failure to do this by the end of the renewal period could result in client applications being unable to connect to the cluster.

Note
For workloads running inside the same Kubernetes cluster and namespace, Secrets can be mounted as a volume so the client Pods construct their keystores and truststores from the current state of the Secrets. For more details on this procedure, see Configuring internal clients to trust the cluster CA.

9.3.3. Manually renewing the CA certificates generated by the Cluster Operator

Cluster and clients CA certificates generated by the Cluster Operator auto-renew at the start of their respective certificate renewal periods. However, you can use the strimzi.io/force-renew annotation to manually renew one or both of these certificates before the certificate renewal period starts. You might do this for security reasons, or if you have changed the renewal or validity periods for the certificates.

A renewed certificate uses the same private key as the old certificate.

Note
If you are using your own CA certificates, the force-renew annotation cannot be used. Instead, follow the procedure for renewing your own CA certificates.
Prerequisites
  • The Cluster Operator is running.

  • A Kafka cluster in which CA certificates and private keys are installed.

Procedure
  1. Apply the strimzi.io/force-renew annotation to the Secret that contains the CA certificate that you want to renew.

    Table 11. Annotation for the Secret that forces renewal of certificates
    Certificate Secret Annotate command

    Cluster CA

    KAFKA-CLUSTER-NAME-cluster-ca-cert

    kubectl annotate secret KAFKA-CLUSTER-NAME-cluster-ca-cert strimzi.io/force-renew=true

    Clients CA

    KAFKA-CLUSTER-NAME-clients-ca-cert

    kubectl annotate secret KAFKA-CLUSTER-NAME-clients-ca-cert strimzi.io/force-renew=true

    At the next reconciliation the Cluster Operator will generate a new CA certificate for the Secret that you annotated. If maintenance time windows are configured, the Cluster Operator will generate the new CA certificate at the first reconciliation within the next maintenance time window.

    Client applications must reload the cluster and clients CA certificates that were renewed by the Cluster Operator.

  2. Check the period the CA certificate is valid:

    For example, using an openssl command:

    kubectl get secret CA-CERTIFICATE-SECRET -o 'jsonpath={.data.CA-CERTIFICATE}' | base64 -d | openssl x509 -subject -issuer -startdate -enddate -noout

    CA-CERTIFICATE-SECRET is the name of the Secret, which is KAFKA-CLUSTER-NAME-cluster-ca-cert for the cluster CA certificate and KAFKA-CLUSTER-NAME-clients-ca-cert for the clients CA certificate.

    CA-CERTIFICATE is the name of the CA certificate, such as jsonpath={.data.ca\.crt}.

    The command returns a notBefore and notAfter date, which is the validity period for the CA certificate.

    For example, for a cluster CA certificate:

    subject=O = io.strimzi, CN = cluster-ca v0
    issuer=O = io.strimzi, CN = cluster-ca v0
    notBefore=Jun 30 09:43:54 2020 GMT
    notAfter=Jun 30 09:43:54 2021 GMT
  3. Delete old certificates from the Secret.

    When components are using the new certificates, older certificates might still be active. Delete the old certificates to remove any potential security risk.

9.3.4. Replacing private keys used by the CA certificates generated by the Cluster Operator

You can replace the private keys used by the cluster CA and clients CA certificates generated by the Cluster Operator. When a private key is replaced, the Cluster Operator generates a new CA certificate for the new private key.

Note
If you are using your own CA certificates, the force-replace annotation cannot be used. Instead, follow the procedure for renewing your own CA certificates.
Prerequisites
  • The Cluster Operator is running.

  • A Kafka cluster in which CA certificates and private keys are installed.

Procedure
  • Apply the strimzi.io/force-replace annotation to the Secret that contains the private key that you want to renew.

    Table 12. Commands for replacing private keys
    Private key for Secret Annotate command

    Cluster CA

    CLUSTER-NAME-cluster-ca

    kubectl annotate secret CLUSTER-NAME-cluster-ca strimzi.io/force-replace=true

    Clients CA

    CLUSTER-NAME-clients-ca

    kubectl annotate secret CLUSTER-NAME-clients-ca strimzi.io/force-replace=true

At the next reconciliation the Cluster Operator will:

  • Generate a new private key for the Secret that you annotated

  • Generate a new CA certificate

If maintenance time windows are configured, the Cluster Operator will generate the new private key and CA certificate at the first reconciliation within the next maintenance time window.

Client applications must reload the cluster and clients CA certificates that were renewed by the Cluster Operator.

9.3.5. Renewing your own CA certificates

This procedure describes how to renew CA certificates and keys you installed yourself, instead of using the certificates generated by the Cluster Operator.

If you are using your own certificates, the Cluster Operator will not renew them automatically. Therefore, it is important that you follow this procedure during the renewal period of the certificate in order to replace CA certificates that will soon expire.

The procedure describes the renewal of CA certificates in PEM format. You can also use certificates in PKCS #12 format.

Prerequisites

These could be generated using an openssl command, such as:

openssl req -x509 -new -days NUMBER-OF-DAYS-VALID --nodes -out ca.crt -keyout ca.key
Procedure
  1. Check the details of the current CA certificates in the Secret:

    kubectl describe secret CA-CERTIFICATE-SECRET

    CA-CERTIFICATE-SECRET is the name of the Secret, which is KAFKA-CLUSTER-NAME-cluster-ca-cert for the cluster CA certificate and KAFKA-CLUSTER-NAME-clients-ca-cert for the clients CA certificate.

  2. Create a directory to contain the existing CA certificates in the secret.

    mkdir new-ca-cert-secret
    cd new-ca-cert-secret
  3. Fetch the secret for each CA certificate you wish to renew:

    kubectl get secret CA-CERTIFICATE-SECRET -o 'jsonpath={.data.CA-CERTIFICATE}' | base64 -d > CA-CERTIFICATE

    Replace CA-CERTIFICATE with the name of each CA certificate.

  4. Rename the old ca.crt file as ca-DATE.crt, where DATE is the certificate expiry date in the format YEAR-MONTH-DAYTHOUR-MINUTE-SECONDZ.

    For example ca-2018-09-27T17-32-00Z.crt.

    mv ca.crt ca-$(date -u -d$(openssl x509 -enddate -noout -in ca.crt | sed 's/.*=//') +'%Y-%m-%dT%H-%M-%SZ').crt
  5. Copy your new CA certificate into the directory, naming it ca.crt:

    cp PATH-TO-NEW-CERTIFICATE ca.crt
  6. Put your CA certificate in the corresponding Secret.

    1. Delete the existing secret:

      kubectl delete secret CA-CERTIFICATE-SECRET

      CA-CERTIFICATE-SECRET is the name of the Secret, as returned in the first step.

      Ignore any "Not Exists" errors.

    2. Recreate the secret:

      kubectl create secret generic CA-CERTIFICATE-SECRET --from-file=.
  7. Delete the directory you created:

    cd ..
    rm -r new-ca-cert-secret
  8. Put your CA key in the corresponding Secret.

    1. Delete the existing secret:

      kubectl delete secret CA-KEY-SECRET

      CA-KEY-SECRET is the name of CA key, which is KAFKA-CLUSTER-NAME-cluster-ca for the cluster CA key and KAFKA-CLUSTER-NAME-clients-ca for the clients CA key.

    2. Recreate the secret with the new CA key:

      kubectl create secret generic CA-KEY-SECRET --from-file=ca.key=CA-KEY-SECRET-FILENAME
  9. Label the secrets with the labels strimzi.io/kind=Kafka and strimzi.io/cluster=KAFKA-CLUSTER-NAME:

    kubectl label secret CA-CERTIFICATE-SECRET strimzi.io/kind=Kafka strimzi.io/cluster=KAFKA-CLUSTER-NAME
    kubectl label secret CA-KEY-SECRET strimzi.io/kind=Kafka strimzi.io/cluster=KAFKA-CLUSTER-NAME

9.4. TLS connections

9.4.1. ZooKeeper communication

Communication between the ZooKeeper nodes on all ports, as well as between clients and ZooKeeper, is encrypted using TLS.

Communication between Kafka brokers and ZooKeeper nodes is also encrypted.

9.4.2. Kafka inter-broker communication

Communication between Kafka brokers is always encrypted using TLS.

Unless the ControlPlaneListener feature gate is enabled, all inter-broker communication goes through an internal listener on port 9091. If you enable the feature gate, traffic from the control plane goes through an internal control plane listener on port 9090. Traffic from the data plane continues to use the existing internal listener on port 9091.

These internal listeners are not available to Kafka clients.

9.4.3. Topic and User Operators

All Operators use encryption for communication with both Kafka and ZooKeeper. In Topic and User Operators, a TLS sidecar is used when communicating with ZooKeeper.

9.4.4. Cruise Control

Cruise Control uses encryption for communication with both Kafka and ZooKeeper. A TLS sidecar is used when communicating with ZooKeeper.

9.4.5. Kafka Client connections

Encrypted or unencrypted communication between Kafka brokers and clients is configured using the tls property for spec.kafka.listeners.

9.5. Configuring internal clients to trust the cluster CA

This procedure describes how to configure a Kafka client that resides inside the Kubernetes cluster — connecting to a TLS listener — to trust the cluster CA certificate.

The easiest way to achieve this for an internal client is to use a volume mount to access the Secrets containing the necessary certificates and keys.

Follow the steps to configure trust certificates that are signed by the cluster CA for Java-based Kafka Producer, Consumer, and Streams APIs.

Choose the steps to follow according to the certificate format of the cluster CA: PKCS #12 (.p12) or PEM (.crt).

The steps describe how to mount the Cluster Secret that verifies the identity of the Kafka cluster to the client pod.

Prerequisites
  • The Cluster Operator must be running.

  • There needs to be a Kafka resource within the Kubernetes cluster.

  • You need a Kafka client application inside the Kubernetes cluster that will connect using TLS, and needs to trust the cluster CA certificate.

  • The client application must be running in the same namespace as the Kafka resource.

Using PKCS #12 format (.p12)
  1. Mount the cluster Secret as a volume when defining the client pod.

    For example:

    kind: Pod
    apiVersion: v1
    metadata:
      name: client-pod
    spec:
      containers:
      - name: client-name
        image: client-name
        volumeMounts:
        - name: secret-volume
          mountPath: /data/p12
        env:
        - name: SECRET_PASSWORD
          valueFrom:
            secretKeyRef:
              name: my-secret
              key: my-password
      volumes:
      - name: secret-volume
        secret:
          secretName: my-cluster-cluster-ca-cert

    Here we’re mounting:

    • The PKCS #12 file into an exact path, which can be configured

    • The password into an environment variable, where it can be used for Java configuration

  2. Configure the Kafka client with the following properties:

    • A security protocol option:

      • security.protocol: SSL when using TLS for encryption (with or without TLS authentication).

      • security.protocol: SASL_SSL when using SCRAM-SHA authentication over TLS.

    • ssl.truststore.location with the truststore location where the certificates were imported.

    • ssl.truststore.password with the password for accessing the truststore.

    • ssl.truststore.type=PKCS12 to identify the truststore type.

Using PEM format (.crt)
  1. Mount the cluster Secret as a volume when defining the client pod.

    For example:

    kind: Pod
    apiVersion: v1
    metadata:
      name: client-pod
    spec:
      containers:
      - name: client-name
        image: client-name
        volumeMounts:
        - name: secret-volume
          mountPath: /data/crt
      volumes:
      - name: secret-volume
        secret:
          secretName: my-cluster-cluster-ca-cert
  2. Use the certificate with clients that use certificates in X.509 format.

9.6. Configuring external clients to trust the cluster CA

This procedure describes how to configure a Kafka client that resides outside the Kubernetes cluster – connecting to an external listener – to trust the cluster CA certificate. Follow this procedure when setting up the client and during the renewal period, when the old clients CA certificate is replaced.

Follow the steps to configure trust certificates that are signed by the cluster CA for Java-based Kafka Producer, Consumer, and Streams APIs.

Choose the steps to follow according to the certificate format of the cluster CA: PKCS #12 (.p12) or PEM (.crt).

The steps describe how to obtain the certificate from the Cluster Secret that verifies the identity of the Kafka cluster.

Important
The <cluster-name>-cluster-ca-cert Secret will contain more than one CA certificate during the CA certificate renewal period. Clients must add all of them to their truststores.
Prerequisites
  • The Cluster Operator must be running.

  • There needs to be a Kafka resource within the Kubernetes cluster.

  • You need a Kafka client application outside the Kubernetes cluster that will connect using TLS, and needs to trust the cluster CA certificate.

Using PKCS #12 format (.p12)
  1. Extract the cluster CA certificate and password from the generated <cluster-name>-cluster-ca-cert Secret.

    kubectl get secret <cluster-name>-cluster-ca-cert -o jsonpath='{.data.ca\.p12}' | base64 -d > ca.p12
    kubectl get secret <cluster-name>-cluster-ca-cert -o jsonpath='{.data.ca\.password}' | base64 -d > ca.password
  2. Configure the Kafka client with the following properties:

    • A security protocol option:

      • security.protocol: SSL when using TLS for encryption (with or without TLS authentication).

      • security.protocol: SASL_SSL when using SCRAM-SHA authentication over TLS.

    • ssl.truststore.location with the truststore location where the certificates were imported.

    • ssl.truststore.password with the password for accessing the truststore. This property can be omitted if it is not needed by the truststore.

    • ssl.truststore.type=PKCS12 to identify the truststore type.

Using PEM format (.crt)
  1. Extract the cluster CA certificate from the generated <cluster-name>-cluster-ca-cert Secret.

    kubectl get secret <cluster-name>-cluster-ca-cert -o jsonpath='{.data.ca\.crt}' | base64 -d > ca.crt
  2. Use the certificate with clients that use certificates in X.509 format.

9.7. Kafka listener certificates

You can provide your own server certificates and private keys for the following types of listeners:

  • Internal TLS listeners for communication within the Kubernetes cluster

  • External listeners (route, loadbalancer, ingress, and nodeport types), which have TLS encryption enabled, for communication between Kafka clients and Kafka brokers

These user-provided certificates are called Kafka listener certificates.

Providing Kafka listener certificates for external listeners allows you to leverage existing security infrastructure, such as your organization’s private CA or a public CA. Kafka clients will connect to Kafka brokers using Kafka listener certificates rather than certificates signed by the cluster CA or clients CA.

You must manually renew Kafka listener certificates when needed.

9.7.1. Providing your own Kafka listener certificates

This procedure shows how to configure a listener to use your own private key and server certificate, called a Kafka listener certificate.

Your client applications should use the CA public key as a trusted certificate in order to verify the identity of the Kafka broker.

Prerequisites
  • A Kubernetes cluster.

  • The Cluster Operator is running.

  • For each listener, a compatible server certificate signed by an external CA.

Procedure
  1. Create a Secret containing your private key and server certificate:

    kubectl create secret generic my-secret --from-file=my-listener-key.key --from-file=my-listener-certificate.crt
  2. Edit the Kafka resource for your cluster. Configure the listener to use your Secret, certificate file, and private key file in the configuration.brokerCertChainAndKey property.

    Example configuration for a loadbalancer external listener with TLS encryption enabled
    # ...
    listeners:
      - name: plain
        port: 9092
        type: internal
        tls: false
      - name: external
        port: 9094
        type: loadbalancer
        tls: true
        authentication:
          type: tls
        configuration:
          brokerCertChainAndKey:
            secretName: my-secret
            certificate: my-listener-certificate.crt
            key: my-listener-key.key
    # ...
    Example configuration for a TLS listener
    # ...
    listeners:
      - name: plain
        port: 9092
        type: internal
        tls: false
      - name: tls
        port: 9093
        type: internal
        tls: true
        authentication:
          type: tls
        configuration:
          brokerCertChainAndKey:
            secretName: my-secret
            certificate: my-listener-certificate.crt
            key: my-listener-key.key
    # ...
  3. Apply the new configuration to create or update the resource:

    kubectl apply -f kafka.yaml

    The Cluster Operator starts a rolling update of the Kafka cluster, which updates the configuration of the listeners.

    Note
    A rolling update is also started if you update a Kafka listener certificate in a Secret that is already used by a TLS or external listener.

9.7.2. Alternative subjects in server certificates for Kafka listeners

In order to use TLS hostname verification with your own Kafka listener certificates, you must use the correct Subject Alternative Names (SANs) for each listener. The certificate SANs must specify hostnames for:

  • All of the Kafka brokers in your cluster

  • The Kafka cluster bootstrap service

You can use wildcard certificates if they are supported by your CA.

TLS listener SAN examples

Use the following examples to help you specify hostnames of the SANs in your certificates for TLS listeners.

Wildcards example
//Kafka brokers
*.<cluster-name>-kafka-brokers
*.<cluster-name>-kafka-brokers.<namespace>.svc

// Bootstrap service
<cluster-name>-kafka-bootstrap
<cluster-name>-kafka-bootstrap.<namespace>.svc
Non-wildcards example
// Kafka brokers
<cluster-name>-kafka-0.<cluster-name>-kafka-brokers
<cluster-name>-kafka-0.<cluster-name>-kafka-brokers.<namespace>.svc
<cluster-name>-kafka-1.<cluster-name>-kafka-brokers
<cluster-name>-kafka-1.<cluster-name>-kafka-brokers.<namespace>.svc
# ...

// Bootstrap service
<cluster-name>-kafka-bootstrap
<cluster-name>-kafka-bootstrap.<namespace>.svc
External listener SAN examples

For external listeners which have TLS encryption enabled, the hostnames you need to specify in certificates depends on the external listener type.

Table 13. SANs for each type of external listener
External listener type In the SANs, specify…​

Route

Addresses of all Kafka broker Routes and the address of the bootstrap Route.

You can use a matching wildcard name.

loadbalancer

Addresses of all Kafka broker loadbalancers and the bootstrap loadbalancer address.

You can use a matching wildcard name.

NodePort

Addresses of all Kubernetes worker nodes that the Kafka broker pods might be scheduled to.

You can use a matching wildcard name.

10. Managing Strimzi

This chapter covers tasks to maintain a deployment of Strimzi.

10.1. Working with custom resources

You can use kubectl commands to retrieve information and perform other operations on Strimzi custom resources.

Using kubectl with the status subresource of a custom resource allows you to get the information about the resource.

10.1.1. Performing kubectl operations on custom resources

Use kubectl commands, such as get, describe, edit, or delete, to perform operations on resource types. For example, kubectl get kafkatopics retrieves a list of all Kafka topics and kubectl get kafkas retrieves all deployed Kafka clusters.

When referencing resource types, you can use both singular and plural names: kubectl get kafkas gets the same results as kubectl get kafka.

You can also use the short name of the resource. Learning short names can save you time when managing Strimzi. The short name for Kafka is k, so you can also run kubectl get k to list all Kafka clusters.

kubectl get k

NAME         DESIRED KAFKA REPLICAS   DESIRED ZK REPLICAS
my-cluster   3                        3
Table 14. Long and short names for each Strimzi resource
Strimzi resource Long name Short name

Kafka

kafka

k

Kafka Topic

kafkatopic

kt

Kafka User

kafkauser

ku

Kafka Connect

kafkaconnect

kc

Kafka Connect S2I

kafkaconnects2i

kcs2i

Kafka Connector

kafkaconnector

kctr

Kafka Mirror Maker

kafkamirrormaker

kmm

Kafka Mirror Maker 2

kafkamirrormaker2

kmm2

Kafka Bridge

kafkabridge

kb

Kafka Rebalance

kafkarebalance

kr

Resource categories

Categories of custom resources can also be used in kubectl commands.

All Strimzi custom resources belong to the category strimzi, so you can use strimzi to get all the Strimzi resources with one command.

For example, running kubectl get strimzi lists all Strimzi custom resources in a given namespace.

kubectl get strimzi

NAME                                   DESIRED KAFKA REPLICAS DESIRED ZK REPLICAS
kafka.kafka.strimzi.io/my-cluster      3                      3

NAME                                   PARTITIONS REPLICATION FACTOR
kafkatopic.kafka.strimzi.io/kafka-apps 3          3

NAME                                   AUTHENTICATION AUTHORIZATION
kafkauser.kafka.strimzi.io/my-user     tls            simple

The kubectl get strimzi -o name command returns all resource types and resource names. The -o name option fetches the output in the type/name format

kubectl get strimzi -o name

kafka.kafka.strimzi.io/my-cluster
kafkatopic.kafka.strimzi.io/kafka-apps
kafkauser.kafka.strimzi.io/my-user

You can combine this strimzi command with other commands. For example, you can pass it into a kubectl delete command to delete all resources in a single command.

kubectl delete $(kubectl get strimzi -o name)

kafka.kafka.strimzi.io "my-cluster" deleted
kafkatopic.kafka.strimzi.io "kafka-apps" deleted
kafkauser.kafka.strimzi.io "my-user" deleted

Deleting all resources in a single operation might be useful, for example, when you are testing new Strimzi features.

Querying the status of sub-resources

There are other values you can pass to the -o option. For example, by using -o yaml you get the output in YAML format. Usng -o json will return it as JSON.

You can see all the options in kubectl get --help.

One of the most useful options is the JSONPath support, which allows you to pass JSONPath expressions to query the Kubernetes API. A JSONPath expression can extract or navigate specific parts of any resource.

For example, you can use the JSONPath expression {.status.listeners[?(@.type=="tls")].bootstrapServers} to get the bootstrap address from the status of the Kafka custom resource and use it in your Kafka clients.

Here, the command finds the bootstrapServers value of the tls listeners.

kubectl get kafka my-cluster -o=jsonpath='{.status.listeners[?(@.type=="tls")].bootstrapServers}{"\n"}'

my-cluster-kafka-bootstrap.myproject.svc:9093

By changing the type condition to @.type=="external" or @.type=="plain" you can also get the address of the other Kafka listeners.

kubectl get kafka my-cluster -o=jsonpath='{.status.listeners[?(@.type=="external")].bootstrapServers}{"\n"}'

192.168.1.247:9094

You can use jsonpath to extract any other property or group of properties from any custom resource.

10.1.2. Strimzi custom resource status information

Several resources have a status property, as described in the following table.

Table 15. Custom resource status properties
Strimzi resource Schema reference Publishes status information on…​

Kafka

KafkaStatus schema reference

The Kafka cluster.

KafkaConnect

KafkaConnectStatus schema reference

The Kafka Connect cluster, if deployed.

KafkaConnectS2I

KafkaConnectS2IStatus schema reference

The Kafka Connect cluster with Source-to-Image support, if deployed.

KafkaConnector

KafkaConnectorStatus schema reference

KafkaConnector resources, if deployed.

KafkaMirrorMaker

KafkaMirrorMakerStatus schema reference

The Kafka MirrorMaker tool, if deployed.

KafkaTopic

KafkaTopicStatus schema reference

Kafka topics in your Kafka cluster.

KafkaUser

KafkaUserStatus schema reference

Kafka users in your Kafka cluster.

KafkaBridge

KafkaBridgeStatus schema reference

The Strimzi Kafka Bridge, if deployed.

The status property of a resource provides information on the resource’s:

  • Current state, in the status.conditions property

  • Last observed generation, in the status.observedGeneration property

The status property also provides resource-specific information. For example:

  • KafkaStatus provides information on listener addresses, and the id of the Kafka cluster.

  • KafkaConnectStatus provides the REST API endpoint for Kafka Connect connectors.

  • KafkaUserStatus provides the user name of the Kafka user and the Secret in which their credentials are stored.

  • KafkaBridgeStatus provides the HTTP address at which external client applications can access the Bridge service.

A resource’s current state is useful for tracking progress related to the resource achieving its desired state, as defined by the spec property. The status conditions provide the time and reason the state of the resource changed and details of events preventing or delaying the operator from realizing the resource’s desired state.

The last observed generation is the generation of the resource that was last reconciled by the Cluster Operator. If the value of observedGeneration is different from the value of metadata.generation, the operator has not yet processed the latest update to the resource. If these values are the same, the status information reflects the most recent changes to the resource.

Strimzi creates and maintains the status of custom resources, periodically evaluating the current state of the custom resource and updating its status accordingly. When performing an update on a custom resource using kubectl edit, for example, its status is not editable. Moreover, changing the status would not affect the configuration of the Kafka cluster.

Here we see the status property specified for a Kafka custom resource.

Kafka custom resource with status
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
spec:
  # ...
status:
  conditions: (1)
  - lastTransitionTime: 2021-07-23T23:46:57+0000
    status: "True"
    type: Ready (2)
  observedGeneration: 4 (3)
  listeners: (4)
  - addresses:
    - host: my-cluster-kafka-bootstrap.myproject.svc
      port: 9092
    type: plain
  - addresses:
    - host: my-cluster-kafka-bootstrap.myproject.svc
      port: 9093
    certificates:
    - |
      -----BEGIN CERTIFICATE-----
      ...
      -----END CERTIFICATE-----
    type: tls
  - addresses:
    - host: 172.29.49.180
      port: 9094
    certificates:
    - |
      -----BEGIN CERTIFICATE-----
      ...
      -----END CERTIFICATE-----
    type: external
  clusterId: CLUSTER-ID (5)
# ...
  1. Status conditions describe criteria related to the status that cannot be deduced from the existing resource information, or are specific to the instance of a resource.

  2. The Ready condition indicates whether the Cluster Operator currently considers the Kafka cluster able to handle traffic.

  3. The observedGeneration indicates the generation of the Kafka custom resource that was last reconciled by the Cluster Operator.

  4. The listeners describe the current Kafka bootstrap addresses by type.

  5. The Kafka cluster id.

    Important
    The address in the custom resource status for external listeners with type nodeport is currently not supported.
Note
The Kafka bootstrap addresses listed in the status do not signify that those endpoints or the Kafka cluster is in a ready state.
Accessing status information

You can access status information for a resource from the command line. For more information, see Finding the status of a custom resource.

10.1.3. Finding the status of a custom resource

This procedure describes how to find the status of a custom resource.

Prerequisites
  • A Kubernetes cluster.

  • The Cluster Operator is running.

Procedure
  • Specify the custom resource and use the -o jsonpath option to apply a standard JSONPath expression to select the status property:

    kubectl get kafka <kafka_resource_name> -o jsonpath='{.status}'

    This expression returns all the status information for the specified custom resource. You can use dot notation, such as status.listeners or status.observedGeneration, to fine-tune the status information you wish to see.

Additional resources

10.2. Pausing reconciliation of custom resources

Sometimes it is useful to pause the reconciliation of custom resources managed by Strimzi Operators, so that you can perform fixes or make updates. If reconciliations are paused, any changes made to custom resources are ignored by the Operators until the pause ends.

If you want to pause reconciliation of a custom resource, set the strimzi.io/pause-reconciliation annotation to true in its configuration. This instructs the appropriate Operator to pause reconciliation of the custom resource. For example, you can apply the annotation to the KafkaConnect resource so that reconciliation by the Cluster Operator is paused.

You can also create a custom resource with the pause annotation enabled. The custom resource is created, but it is ignored.

Prerequisites
  • The Strimzi Operator that manages the custom resource is running.

Procedure
  1. Annotate the custom resource in Kubernetes, setting pause-reconciliation to true:

    kubectl annotate KIND-OF-CUSTOM-RESOURCE NAME-OF-CUSTOM-RESOURCE strimzi.io/pause-reconciliation="true"

    For example, for the KafkaConnect custom resource:

    kubectl annotate KafkaConnect my-connect strimzi.io/pause-reconciliation="true"
  2. Check that the status conditions of the custom resource show a change to ReconciliationPaused:

    kubectl describe KIND-OF-CUSTOM-RESOURCE NAME-OF-CUSTOM-RESOURCE

    The type condition changes to ReconciliationPaused at the lastTransitionTime.

    Example custom resource with a paused reconciliation condition type
    apiVersion: kafka.strimzi.io/v1beta2
    kind: KafkaConnect
    metadata:
      annotations:
        strimzi.io/pause-reconciliation: "true"
        strimzi.io/use-connector-resources: "true"
      creationTimestamp: 2021-03-12T10:47:11Z
      #...
    spec:
      # ...
    status:
      conditions:
      - lastTransitionTime: 2021-03-12T10:47:41.689249Z
        status: "True"
        type: ReconciliationPaused
Resuming from pause
  • To resume reconciliation, you can set the annotation to false, or remove the annotation.

10.3. Manually starting rolling updates of Kafka and ZooKeeper clusters

Strimzi supports the use of annotations on StatefulSet and Pod resources to manually trigger a rolling update of Kafka and ZooKeeper clusters through the Cluster Operator. Rolling updates restart the pods of the resource with new ones.

Manually performing a rolling update on a specific pod or set of pods from the same StatefulSet is usually only required in exceptional circumstances. However, rather than deleting the pods directly, if you perform the rolling update through the Cluster Operator you ensure that:

  • The manual deletion of the pod does not conflict with simultaneous Cluster Operator operations, such as deleting other pods in parallel.

  • The Cluster Operator logic handles the Kafka configuration specifications, such as the number of in-sync replicas.

10.3.1. Prerequisites

To perform a manual rolling update, you need a running Cluster Operator and Kafka cluster.

See the Deploying and Upgrading Strimzi guide for instructions on running a:

10.3.2. Performing a rolling update using a StatefulSet annotation

This procedure describes how to manually trigger a rolling update of an existing Kafka cluster or ZooKeeper cluster using a Kubernetes StatefulSet annotation.

Procedure
  1. Find the name of the StatefulSet that controls the Kafka or ZooKeeper pods you want to manually update.

    For example, if your Kafka cluster is named my-cluster, the corresponding StatefulSet names are my-cluster-kafka and my-cluster-zookeeper.

  2. Annotate the StatefulSet resource in Kubernetes.

    Use kubectl annotate:

    kubectl annotate statefulset cluster-name-kafka strimzi.io/manual-rolling-update=true
    
    kubectl annotate statefulset cluster-name-zookeeper strimzi.io/manual-rolling-update=true
  3. Wait for the next reconciliation to occur (every two minutes by default). A rolling update of all pods within the annotated StatefulSet is triggered, as long as the annotation was detected by the reconciliation process. When the rolling update of all the pods is complete, the annotation is removed from the StatefulSet.

10.3.3. Performing a rolling update using a Pod annotation

This procedure describes how to manually trigger a rolling update of an existing Kafka cluster or ZooKeeper cluster using a Kubernetes Pod annotation. When multiple pods from the same StatefulSet are annotated, consecutive rolling updates are performed within the same reconciliation run.

Procedure
  1. Find the name of the Kafka or ZooKeeper Pod you want to manually update.

    For example, if your Kafka cluster is named my-cluster, the corresponding Pod names are my-cluster-kafka-index and my-cluster-zookeeper-index. The index starts at zero and ends at the total number of replicas.

  2. Annotate the Pod resource in Kubernetes.

    Use kubectl annotate:

    kubectl annotate pod cluster-name-kafka-index strimzi.io/manual-rolling-update=true
    
    kubectl annotate pod cluster-name-zookeeper-index strimzi.io/manual-rolling-update=true
  3. Wait for the next reconciliation to occur (every two minutes by default). A rolling update of the annotated Pod is triggered, as long as the annotation was detected by the reconciliation process. When the rolling update of a pod is complete, the annotation is removed from the Pod.

10.4. Discovering services using labels and annotations

Service discovery makes it easier for client applications running in the same Kubernetes cluster as Strimzi to interact with a Kafka cluster.

A service discovery label and annotation is generated for services used to access the Kafka cluster:

  • Internal Kafka bootstrap service

  • HTTP Bridge service

The label helps to make the service discoverable, and the annotation provides connection details that a client application can use to make the connection.

The service discovery label, strimzi.io/discovery, is set as true for the Service resources. The service discovery annotation has the same key, providing connection details in JSON format for each service.

Example internal Kafka bootstrap service

apiVersion: v1
kind: Service
metadata:
  annotations:
    strimzi.io/discovery: |-
      [ {
        "port" : 9092,
        "tls" : false,
        "protocol" : "kafka",
        "auth" : "scram-sha-512"
      }, {
        "port" : 9093,
        "tls" : true,
        "protocol" : "kafka",
        "auth" : "tls"
      } ]
  labels:
    strimzi.io/cluster: my-cluster
    strimzi.io/discovery: "true"
    strimzi.io/kind: Kafka
    strimzi.io/name: my-cluster-kafka-bootstrap
  name: my-cluster-kafka-bootstrap
spec:
  #...

Example HTTP Bridge service

apiVersion: v1
kind: Service
metadata:
  annotations:
    strimzi.io/discovery: |-
      [ {
        "port" : 8080,
        "tls" : false,
        "auth" : "none",
        "protocol" : "http"
      } ]
  labels:
    strimzi.io/cluster: my-bridge
    strimzi.io/discovery: "true"
    strimzi.io/kind: KafkaBridge
    strimzi.io/name: my-bridge-bridge-service

10.4.1. Returning connection details on services

You can find the services by specifying the discovery label when fetching services from the command line or a corresponding API call.

kubectl get service -l strimzi.io/discovery=true

The connection details are returned when retrieving the service discovery label.

10.5. Recovering a cluster from persistent volumes

You can recover a Kafka cluster from persistent volumes (PVs) if they are still present.

You might want to do this, for example, after:

  • A namespace was deleted unintentionally

  • A whole Kubernetes cluster is lost, but the PVs remain in the infrastructure

10.5.1. Recovery from namespace deletion

Recovery from namespace deletion is possible because of the relationship between persistent volumes and namespaces. A PersistentVolume (PV) is a storage resource that lives outside of a namespace. A PV is mounted into a Kafka pod using a PersistentVolumeClaim (PVC), which lives inside a namespace.

The reclaim policy for a PV tells a cluster how to act when a namespace is deleted. If the reclaim policy is set as:

  • Delete (default), PVs are deleted when PVCs are deleted within a namespace

  • Retain, PVs are not deleted when a namespace is deleted

To ensure that you can recover from a PV if a namespace is deleted unintentionally, the policy must be reset from Delete to Retain in the PV specification using the persistentVolumeReclaimPolicy property:

apiVersion: v1
kind: PersistentVolume
# ...
spec:
  # ...
  persistentVolumeReclaimPolicy: Retain

Alternatively, PVs can inherit the reclaim policy of an associated storage class. Storage classes are used for dynamic volume allocation.

By configuring the reclaimPolicy property for the storage class, PVs that use the storage class are created with the appropriate reclaim policy. The storage class is configured for the PV using the storageClassName property.

apiVersion: v1
kind: StorageClass
metadata:
  name: gp2-retain
parameters:
  # ...
# ...
reclaimPolicy: Retain
apiVersion: v1
kind: PersistentVolume
# ...
spec:
  # ...
  storageClassName: gp2-retain
Note
If you are using Retain as the reclaim policy, but you want to delete an entire cluster, you need to delete the PVs manually. Otherwise they will not be deleted, and may cause unnecessary expenditure on resources.

10.5.2. Recovery from loss of a Kubernetes cluster

When a cluster is lost, you can use the data from disks/volumes to recover the cluster if they were preserved within the infrastructure. The recovery procedure is the same as with namespace deletion, assuming PVs can be recovered and they were created manually.

10.5.3. Recovering a deleted cluster from persistent volumes

This procedure describes how to recover a deleted cluster from persistent volumes (PVs).

In this situation, the Topic Operator identifies that topics exist in Kafka, but the KafkaTopic resources do not exist.

When you get to the step to recreate your cluster, you have two options:

  1. Use Option 1 when you can recover all KafkaTopic resources.

    The KafkaTopic resources must therefore be recovered before the cluster is started so that the corresponding topics are not deleted by the Topic Operator.

  2. Use Option 2 when you are unable to recover all KafkaTopic resources.

    In this case, you deploy your cluster without the Topic Operator, delete the Topic Operator topic store metadata, and then redeploy the Kafka cluster with the Topic Operator so it can recreate the KafkaTopic resources from the corresponding topics.

Note
If the Topic Operator is not deployed, you only need to recover the PersistentVolumeClaim (PVC) resources.
Before you begin

In this procedure, it is essential that PVs are mounted into the correct PVC to avoid data corruption. A volumeName is specified for the PVC and this must match the name of the PV.

For more information, see:

Note
The procedure does not include recovery of KafkaUser resources, which must be recreated manually. If passwords and certificates need to be retained, secrets must be recreated before creating the KafkaUser resources.
Procedure
  1. Check information on the PVs in the cluster:

    kubectl get pv

    Information is presented for PVs with data.

    Example output showing columns important to this procedure:

    NAME                                         RECLAIMPOLICY CLAIM
    pvc-5e9c5c7f-3317-11ea-a650-06e1eadd9a4c ... Retain ...    myproject/data-my-cluster-zookeeper-1
    pvc-5e9cc72d-3317-11ea-97b0-0aef8816c7ea ... Retain ...    myproject/data-my-cluster-zookeeper-0
    pvc-5ead43d1-3317-11ea-97b0-0aef8816c7ea ... Retain ...    myproject/data-my-cluster-zookeeper-2
    pvc-7e1f67f9-3317-11ea-a650-06e1eadd9a4c ... Retain ...    myproject/data-0-my-cluster-kafka-0
    pvc-7e21042e-3317-11ea-9786-02deaf9aa87e ... Retain ...    myproject/data-0-my-cluster-kafka-1
    pvc-7e226978-3317-11ea-97b0-0aef8816c7ea ... Retain ...    myproject/data-0-my-cluster-kafka-2
    • NAME shows the name of each PV.

    • RECLAIM POLICY shows that PVs are retained.

    • CLAIM shows the link to the original PVCs.

  2. Recreate the original namespace:

    kubectl create namespace myproject
  3. Recreate the original PVC resource specifications, linking the PVCs to the appropriate PV:

    For example:

    apiVersion: v1
    kind: PersistentVolumeClaim
    metadata:
      name: data-0-my-cluster-kafka-0
    spec:
      accessModes:
      - ReadWriteOnce
      resources:
        requests:
          storage: 100Gi
      storageClassName: gp2-retain
      volumeMode: Filesystem
      volumeName: pvc-7e1f67f9-3317-11ea-a650-06e1eadd9a4c
  4. Edit the PV specifications to delete the claimRef properties that bound the original PVC.

    For example:

    apiVersion: v1
    kind: PersistentVolume
    metadata:
      annotations:
        kubernetes.io/createdby: aws-ebs-dynamic-provisioner
        pv.kubernetes.io/bound-by-controller: "yes"
        pv.kubernetes.io/provisioned-by: kubernetes.io/aws-ebs
      creationTimestamp: "<date>"
      finalizers:
      - kubernetes.io/pv-protection
      labels:
        failure-domain.beta.kubernetes.io/region: eu-west-1
        failure-domain.beta.kubernetes.io/zone: eu-west-1c
      name: pvc-7e226978-3317-11ea-97b0-0aef8816c7ea
      resourceVersion: "39431"
      selfLink: /api/v1/persistentvolumes/pvc-7e226978-3317-11ea-97b0-0aef8816c7ea
      uid: 7efe6b0d-3317-11ea-a650-06e1eadd9a4c
    spec:
      accessModes:
      - ReadWriteOnce
      awsElasticBlockStore:
        fsType: xfs
        volumeID: aws://eu-west-1c/vol-09db3141656d1c258
      capacity:
        storage: 100Gi
      claimRef:
        apiVersion: v1
        kind: PersistentVolumeClaim
        name: data-0-my-cluster-kafka-2
        namespace: myproject
        resourceVersion: "39113"
        uid: 54be1c60-3319-11ea-97b0-0aef8816c7ea
      nodeAffinity:
        required:
          nodeSelectorTerms:
          - matchExpressions:
            - key: failure-domain.beta.kubernetes.io/zone
              operator: In
              values:
              - eu-west-1c
            - key: failure-domain.beta.kubernetes.io/region
              operator: In
              values:
              - eu-west-1
      persistentVolumeReclaimPolicy: Retain
      storageClassName: gp2-retain
      volumeMode: Filesystem

    In the example, the following properties are deleted:

    claimRef:
      apiVersion: v1
      kind: PersistentVolumeClaim
      name: data-0-my-cluster-kafka-2
      namespace: myproject
      resourceVersion: "39113"
      uid: 54be1c60-3319-11ea-97b0-0aef8816c7ea
  5. Deploy the Cluster Operator.

    kubectl create -f install/cluster-operator -n my-project
  6. Recreate your cluster.

    Follow the steps depending on whether or not you have all the KafkaTopic resources needed to recreate your cluster.

    Option 1: If you have all the KafkaTopic resources that existed before you lost your cluster, including internal topics such as committed offsets from __consumer_offsets:

    1. Recreate all KafkaTopic resources.

      It is essential that you recreate the resources before deploying the cluster, or the Topic Operator will delete the topics.

    2. Deploy the Kafka cluster.

      For example:

      kubectl apply -f kafka.yaml

    Option 2: If you do not have all the KafkaTopic resources that existed before you lost your cluster:

    1. Deploy the Kafka cluster, as with the first option, but without the Topic Operator by removing the topicOperator property from the Kafka resource before deploying.

      If you include the Topic Operator in the deployment, the Topic Operator will delete all the topics.

    2. Delete the internal topic store topics from the Kafka cluster:

      kubectl run kafka-admin -ti --image=quay.io/strimzi/kafka:0.24.0-kafka-2.8.0 --rm=true --restart=Never -- ./bin/kafka-topics.sh --bootstrap-server localhost:9092 --topic __strimzi-topic-operator-kstreams-topic-store-changelog --delete && ./bin/kafka-topics.sh --bootstrap-server localhost:9092 --topic __strimzi_store_topic --delete

      The command must correspond to the type of listener and authentication used to access the Kafka cluster.

    3. Enable the Topic Operator by redeploying the Kafka cluster with the topicOperator property to recreate the KafkaTopic resources.

      For example:

      apiVersion: kafka.strimzi.io/v1beta2
      kind: Kafka
      metadata:
        name: my-cluster
      spec:
        #...
        entityOperator:
          topicOperator: {} (1)
          #...
    1. Here we show the default configuration, which has no additional properties. You specify the required configuration using the properties described in EntityTopicOperatorSpec schema reference.

  7. Verify the recovery by listing the KafkaTopic resources:

    kubectl get KafkaTopic

10.6. Tuning Kafka configuration

Use configuration properties to optimize the performance of Kafka brokers, producers and consumers.

A minimum set of configuration properties is required, but you can add or adjust properties to change how producers and consumers interact with Kafka brokers. For example, you can tune latency and throughput of messages so that clients can respond to data in real time.

You might start by analyzing metrics to gauge where to make your initial configurations, then make incremental changes and further comparisons of metrics until you have the configuration you need.

Additional resources

10.6.1. Kafka broker configuration tuning

Use configuration properties to optimize the performance of Kafka brokers. You can use standard Kafka broker configuration options, except for properties managed directly by Strimzi.

Basic broker configuration

Certain broker configuration options are managed directly by Strimzi, driven by the Kafka custom resource specification:

  • broker.id is the ID of the Kafka broker

  • log.dirs are the directories for log data

  • zookeeper.connect is the configuration to connect Kafka with ZooKeeper

  • listener exposes the Kafka cluster to clients

  • authorization mechanisms allow or decline actions executed by users

  • authentication mechanisms prove the identity of users requiring access to Kafka

Broker IDs start from 0 (zero) and correspond to the number of broker replicas. Log directories are mounted to /var/lib/kafka/data/kafka-logIDX based on the spec.kafka.storage configuration in the Kafka custom resource. IDX is the Kafka broker pod index.

As such, you cannot configure these options through the config property of the Kafka custom resource. For a list of exclusions, see the KafkaClusterSpec schema reference.

However, a typical broker configuration will include settings for properties related to topics, threads and logs.

Basic broker configuration properties
# ...
num.partitions=1
default.replication.factor=3
offsets.topic.replication.factor=3
transaction.state.log.replication.factor=3
transaction.state.log.min.isr=2
log.retention.hours=168
log.segment.bytes=1073741824
log.retention.check.interval.ms=300000
num.network.threads=3
num.io.threads=8
num.recovery.threads.per.data.dir=1
socket.send.buffer.bytes=102400
socket.receive.buffer.bytes=102400
socket.request.max.bytes=104857600
group.initial.rebalance.delay.ms=0
zookeeper.connection.timeout.ms=6000
# ...
Replicating topics for high availability

Basic topic properties set the default number of partitions and replication factor for topics, which will apply to topics that are created without these properties being explicitly set, including when topics are created automatically.

# ...
num.partitions=1
auto.create.topics.enable=false
default.replication.factor=3
min.insync.replicas=2
replica.fetch.max.bytes=1048576
# ...

The auto.create.topics.enable property is enabled by default so that topics that do not already exist are created automatically when needed by producers and consumers. If you are using automatic topic creation, you can set the default number of partitions for topics using num.partitions. Generally, however, this property is disabled so that more control is provided over topics through explicit topic creation For example, you can use the Strimzi KafkaTopic resource or applications to create topics.

For high availability environments, it is advisable to increase the replication factor to at least 3 for topics and set the minimum number of in-sync replicas required to 1 less than the replication factor. For topics created using the KafkaTopic resource, the replication factor is set using spec.replicas.

For data durability, you should also set min.insync.replicas in your topic configuration and message delivery acknowledgments using acks=all in your producer configuration.

Use replica.fetch.max.bytes to set the maximum size, in bytes, of messages fetched by each follower that replicates the leader partition. Change this value according to the average message size and throughput. When considering the total memory allocation required for read/write buffering, the memory available must also be able to accommodate the maximum replicated message size when multiplied by all followers.

The delete.topic.enable property is enabled by default to allow topics to be deleted. In a production environment, you should disable this property to avoid accidental topic deletion, resulting in data loss. You can, however, temporarily enable it and delete topics and then disable it again. If delete.topic.enable is enabled, you can delete topics using the KafkaTopic resource.

# ...
auto.create.topics.enable=false
delete.topic.enable=true
# ...
Internal topic settings for transactions and commits

If you are using transactions to enable atomic writes to partitions from producers, the state of the transactions is stored in the internal __transaction_state topic. By default, the brokers are configured with a replication factor of 3 and a minimum of 2 in-sync replicas for this topic, which means that a minimum of three brokers are required in your Kafka cluster.

# ...
transaction.state.log.replication.factor=3
transaction.state.log.min.isr=2
# ...

Similarly, the internal __consumer_offsets topic, which stores consumer state, has default settings for the number of partitions and replication factor.

# ...
offsets.topic.num.partitions=50
offsets.topic.replication.factor=3
# ...

Do not reduce these settings in production. You can increase the settings in a production environment. As an exception, you might want to reduce the settings in a single-broker test environment.

Improving request handling throughput by increasing I/O threads

Network threads handle requests to the Kafka cluster, such as produce and fetch requests from client applications. Produce requests are placed in a request queue. Responses are placed in a response queue.

The number of network threads should reflect the replication factor and the levels of activity from client producers and consumers interacting with the Kafka cluster. If you are going to have a lot of requests, you can increase the number of threads, using the amount of time threads are idle to determine when to add more threads.

To reduce congestion and regulate the request traffic, you can limit the number of requests allowed in the request queue before the network thread is blocked.

I/O threads pick up requests from the request queue to process them. Adding more threads can improve throughput, but the number of CPU cores and disk bandwidth imposes a practical upper limit. At a minimum, the number of I/O threads should equal the number of storage volumes.

# ...
num.network.threads=3 (1)
queued.max.requests=500 (2)
num.io.threads=8 (3)
num.recovery.threads.per.data.dir=1 (4)
# ...
  1. The number of network threads for the Kafka cluster.

  2. The number of requests allowed in the request queue.

  3. The number of I/O threads for a Kafka broker.

  4. The number of threads used for log loading at startup and flushing at shutdown.

Configuration updates to the thread pools for all brokers might occur dynamically at the cluster level. These updates are restricted to between half the current size and twice the current size.

Note
Kafka broker metrics can help with working out the number of threads required. For example, metrics for the average time network threads are idle (kafka.network:type=SocketServer,name=NetworkProcessorAvgIdlePercent) indicate the percentage of resources used. If there is 0% idle time, all resources are in use, which means that adding more threads might be beneficial.

If threads are slow or limited due to the number of disks, you can try increasing the size of the buffers for network requests to improve throughput:

# ...
replica.socket.receive.buffer.bytes=65536
# ...

And also increase the maximum number of bytes Kafka can receive:

# ...
socket.request.max.bytes=104857600
# ...
Increasing bandwidth for high latency connections

Kafka batches data to achieve reasonable throughput over high-latency connections from Kafka to clients, such as connections between datacenters. However, if high latency is a problem, you can increase the size of the buffers for sending and receiving messages.

# ...
socket.send.buffer.bytes=1048576
socket.receive.buffer.bytes=1048576
# ...

You can estimate the optimal size of your buffers using a bandwidth-delay product calculation, which multiplies the maximum bandwidth of the link (in bytes/s) with the round-trip delay (in seconds) to give an estimate of how large a buffer is required to sustain maximum throughput.

Managing logs with data retention policies

Kafka uses logs to store message data. Logs are a series of segments associated with various indexes. New messages are written to an active segment, and never subsequently modified. Segments are read when serving fetch requests from consumers. Periodically, the active segment is rolled to become read-only and a new active segment is created to replace it. There is only a single segment active at a time. Older segments are retained until they are eligible for deletion.

Configuration at the broker level sets the maximum size in bytes of a log segment and the amount of time in milliseconds before an active segment is rolled:

# ...
log.segment.bytes=1073741824
log.roll.ms=604800000
# ...

You can override these settings at the topic level using segment.bytes and segment.ms. Whether you need to lower or raise these values depends on the policy for segment deletion. A larger size means the active segment contains more messages and is rolled less often. Segments also become eligible for deletion less often.

You can set time-based or size-based log retention and cleanup policies so that logs are kept manageable. Depending on your requirements, you can use log retention configuration to delete old segments. If log retention policies are used, non-active log segments are removed when retention limits are reached. Deleting old segments bounds the storage space required for the log so you do not exceed disk capacity.

For time-based log retention, you set a retention period based on hours, minutes and milliseconds. The retention period is based on the time messages were appended to the segment.

The milliseconds configuration has priority over minutes, which has priority over hours. The minutes and milliseconds configuration is null by default, but the three options provide a substantial level of control over the data you wish to retain. Preference should be given to the milliseconds configuration, as it is the only one of the three properties that is dynamically updateable.

# ...
log.retention.ms=1680000
# ...

If log.retention.ms is set to -1, no time limit is applied to log retention, so all logs are retained. Disk usage should always be monitored, but the -1 setting is not generally recommended as it can lead to issues with full disks, which can be hard to rectify.

For size-based log retention, you set a maximum log size (of all segments in the log) in bytes:

# ...
log.retention.bytes=1073741824
# ...

In other words, a log will typically have approximately log.retention.bytes/log.segment.bytes segments once it reaches a steady state. When the maximum log size is reached, older segments are removed.

A potential issue with using a maximum log size is that it does not take into account the time messages were appended to a segment. You can use time-based and size-based log retention for your cleanup policy to get the balance you need. Whichever threshold is reached first triggers the cleanup.

If you wish to add a time delay before a segment file is deleted from the system, you can add the delay using log.segment.delete.delay.ms for all topics at the broker level or file.delete.delay.ms for specific topics in the topic configuration.

# ...
log.segment.delete.delay.ms=60000
# ...
Removing log data with cleanup policies

The method of removing older log data is determined by the log cleaner configuration.

The log cleaner is enabled for the broker by default:

# ...
log.cleaner.enable=true
# ...

You can set the cleanup policy at the topic or broker level. Broker-level configuration is the default for topics that do not have policy set.

You can set policy to delete logs, compact logs, or do both:

# ...
log.cleanup.policy=compact,delete
# ...

The delete policy corresponds to managing logs with data retention policies. It is suitable when data does not need to be retained forever. The compact policy guarantees to keep the most recent message for each message key. Log compaction is suitable where message values are changeable, and you want to retain the latest update.

If cleanup policy is set to delete logs, older segments are deleted based on log retention limits. Otherwise, if the log cleaner is not enabled, and there are no log retention limits, the log will continue to grow.

If cleanup policy is set for log compaction, the head of the log operates as a standard Kafka log, with writes for new messages appended in order. In the tail of a compacted log, where the log cleaner operates, records will be deleted if another record with the same key occurs later in the log. Messages with null values are also deleted. If you’re not using keys, you can’t use compaction because keys are needed to identify related messages. While Kafka guarantees that the latest messages for each key will be retained, it does not guarantee that the whole compacted log will not contain duplicates.

Image of compaction showing key value writes
Figure 6. Log showing key value writes with offset positions before compaction

Using keys to identify messages, Kafka compaction keeps the latest message (with the highest offset) for a specific message key, eventually discarding earlier messages that have the same key. In other words, the message in its latest state is always available and any out-of-date records of that particular message are eventually removed when the log cleaner runs. You can restore a message back to a previous state.

Records retain their original offsets even when surrounding records get deleted. Consequently, the tail can have non-contiguous offsets. When consuming an offset that’s no longer available in the tail, the record with the next higher offset is found.

Image of compaction after log cleanup
Figure 7. Log after compaction

If you choose only a compact policy, your log can still become arbitrarily large. In which case, you can set policy to compact and delete logs. If you choose to compact and delete, first the log data is compacted, removing records with a key in the head of the log. After which, data that falls before the log retention threshold is deleted.

Image of compaction with retention point
Figure 8. Log retention point and compaction point

You set the frequency the log is checked for cleanup in milliseconds:

# ...
log.retention.check.interval.ms=300000
# ...

Adjust the log retention check interval in relation to the log retention settings. Smaller retention sizes might require more frequent checks.

The frequency of cleanup should be often enough to manage the disk space, but not so often it affects performance on a topic.

You can also set a time in milliseconds to put the cleaner on standby if there are no logs to clean:

# ...
log.cleaner.backoff.ms=15000
# ...

If you choose to delete older log data, you can set a period in milliseconds to retain the deleted data before it is purged:

# ...
log.cleaner.delete.retention.ms=86400000
# ...

The deleted data retention period gives time to notice the data is gone before it is irretrievably deleted.

To delete all messages related to a specific key, a producer can send a tombstone message. A tombstone has a null value and acts as a marker to tell a consumer the value is deleted. After compaction, only the tombstone is retained, which must be for a long enough period for the consumer to know that the message is deleted. When older messages are deleted, having no value, the tombstone key is also deleted from the partition.

Managing disk utilization

There are many other configuration settings related to log cleanup, but of particular importance is memory allocation.

The deduplication property specifies the total memory for cleanup across all log cleaner threads. You can set an upper limit on the percentage of memory used through the buffer load factor.

# ...
log.cleaner.dedupe.buffer.size=134217728
log.cleaner.io.buffer.load.factor=0.9
# ...

Each log entry uses exactly 24 bytes, so you can work out how many log entries the buffer can handle in a single run and adjust the setting accordingly.

If possible, consider increasing the number of log cleaner threads if you are looking to reduce the log cleaning time:

# ...
log.cleaner.threads=8
# ...

If you are experiencing issues with 100% disk bandwidth usage, you can throttle the log cleaner I/O so that the sum of the read/write operations is less than a specified double value based on the capabilities of the disks performing the operations:

# ...
log.cleaner.io.max.bytes.per.second=1.7976931348623157E308
# ...
Handling large message sizes

The default batch size for messages is 1MB, which is optimal for maximum throughput in most use cases. Kafka can accommodate larger batches at a reduced throughput, assuming adequate disk capacity.

Large message sizes are handled in four ways:

  1. Producer-side message compression writes compressed messages to the log.

  2. Reference-based messaging sends only a reference to data stored in some other system in the message’s value.

  3. Inline messaging splits messages into chunks that use the same key, which are then combined on output using a stream-processor like Kafka Streams.

  4. Broker and producer/consumer client application configuration built to handle larger message sizes.

The reference-based messaging and message compression options are recommended and cover most situations. With any of these options, care must be take to avoid introducing performance issues.

Producer-side compression

For producer configuration, you specify a compression.type, such as Gzip, which is then applied to batches of data generated by the producer. Using the broker configuration compression.type=producer, the broker retains whatever compression the producer used. Whenever producer and topic compression do not match, the broker has to compress batches again prior to appending them to the log, which impacts broker performance.

Compression also adds additional processing overhead on the producer and decompression overhead on the consumer, but includes more data in a batch, so is often beneficial to throughput when message data compresses well.

Combine producer-side compression with fine-tuning of the batch size to facilitate optimum throughput. Using metrics helps to gauge the average batch size needed.

Reference-based messaging

Reference-based messaging is useful for data replication when you do not know how big a message will be. The external data store must be fast, durable, and highly available for this configuration to work. Data is written to the data store and a reference to the data is returned. The producer sends a message containing the reference to Kafka. The consumer gets the reference from the message and uses it to fetch the data from the data store.

Image of reference-based messaging flow
Figure 9. Reference-based messaging flow

As the message passing requires more trips, end-to-end latency will increase. Another significant drawback of this approach is there is no automatic clean up of the data in the external system when the Kafka message gets cleaned up. A hybrid approach would be to only send large messages to the data store and process standard-sized messages directly.

Inline messaging

Inline messaging is complex, but it does not have the overhead of depending on external systems like reference-based messaging.

The producing client application has to serialize and then chunk the data if the message is too big. The producer then uses the Kafka ByteArraySerializer or similar to serialize each chunk again before sending it. The consumer tracks messages and buffers chunks until it has a complete message. The consuming client application receives the chunks, which are assembled before deserialization. Complete messages are delivered to the rest of the consuming application in order according to the offset of the first or last chunk for each set of chunked messages. Successful delivery of the complete message is checked against offset metadata to avoid duplicates during a rebalance.

Image of inline messaging flow
Figure 10. Inline messaging flow

Inline messaging has a performance overhead on the consumer side because of the buffering required, particularly when handling a series of large messages in parallel. The chunks of large messages can become interleaved, so that it is not always possible to commit when all the chunks of a message have been consumed if the chunks of another large message in the buffer are incomplete. For this reason, the buffering is usually supported by persisting message chunks or by implementing commit logic.

Configuration to handle larger messages

If larger messages cannot be avoided, and to avoid blocks at any point of the message flow, you can increase message limits. To do this, configure message.max.bytes at the topic level to set the maximum record batch size for individual topics. If you set message.max.bytes at the broker level, larger messages are allowed for all topics.

The broker will reject any message that is greater than the limit set with message.max.bytes. The buffer size for the producers (max.request.size) and consumers (message.max.bytes) must be able to accommodate the larger messages.

Controlling the log flush of message data

Log flush properties control the periodic writes of cached message data to disk. The scheduler specifies the frequency of checks on the log cache in milliseconds:

# ...
log.flush.scheduler.interval.ms=2000
# ...

You can control the frequency of the flush based on the maximum amount of time that a message is kept in-memory and the maximum number of messages in the log before writing to disk:

# ...
log.flush.interval.ms=50000
log.flush.interval.messages=100000
# ...

The wait between flushes includes the time to make the check and the specified interval before the flush is carried out. Increasing the frequency of flushes can affect throughput.

Generally, the recommendation is to not set explicit flush thresholds and let the operating system perform background flush using its default settings. Partition replication provides greater data durability than writes to any single disk as a failed broker can recover from its in-sync replicas.

If you are using application flush management, setting lower flush thresholds might be appropriate if you are using faster disks.

Partition rebalancing for availability

Partitions can be replicated across brokers for fault tolerance. For a given partition, one broker is elected leader and handles all produce requests (writes to the log). Partition followers on other brokers replicate the partition data of the partition leader for data reliability in the event of the leader failing.

Followers do not normally serve clients, though rack configuration allows a consumer to consume messages from the closest replica when a Kafka cluster spans multiple datacenters. Followers operate only to replicate messages from the partition leader and allow recovery should the leader fail. Recovery requires an in-sync follower. Followers stay in sync by sending fetch requests to the leader, which returns messages to the follower in order. The follower is considered to be in sync if it has caught up with the most recently committed message on the leader. The leader checks this by looking at the last offset requested by the follower. An out-of-sync follower is usually not eligible as a leader should the current leader fail, unless unclean leader election is allowed.

You can adjust the lag time before a follower is considered out of sync:

# ...
replica.lag.time.max.ms=30000
# ...

Lag time puts an upper limit on the time to replicate a message to all in-sync replicas and how long a producer has to wait for an acknowledgment. If a follower fails to make a fetch request and catch up with the latest message within the specified lag time, it is removed from in-sync replicas. You can reduce the lag time to detect failed replicas sooner, but by doing so you might increase the number of followers that fall out of sync needlessly. The right lag time value depends on both network latency and broker disk bandwidth.

When a leader partition is no longer available, one of the in-sync replicas is chosen as the new leader. The first broker in a partition’s list of replicas is known as the preferred leader. By default, Kafka is enabled for automatic partition leader rebalancing based on a periodic check of leader distribution. That is, Kafka checks to see if the preferred leader is the current leader. A rebalance ensures that leaders are evenly distributed across brokers and brokers are not overloaded.

You can use Cruise Control for Strimzi to figure out replica assignments to brokers that balance load evenly across the cluster. Its calculation takes into account the differing load experienced by leaders and followers. A failed leader affects the balance of a Kafka cluster because the remaining brokers get the extra work of leading additional partitions.

For the assignment found by Cruise Control to actually be balanced it is necessary that partitions are lead by the preferred leader. Kafka can automatically ensure that the preferred leader is being used (where possible), changing the current leader if necessary. This ensures that the cluster remains in the balanced state found by Cruise Control.

You can control the frequency, in seconds, of the rebalance check and the maximum percentage of imbalance allowed for a broker before a rebalance is triggered.

#...
auto.leader.rebalance.enable=true
leader.imbalance.check.interval.seconds=300
leader.imbalance.per.broker.percentage=10
#...

The percentage leader imbalance for a broker is the ratio between the current number of partitions for which the broker is the current leader and the number of partitions for which it is the preferred leader. You can set the percentage to zero to ensure that preferred leaders are always elected, assuming they are in sync.

If the checks for rebalances need more control, you can disable automated rebalances. You can then choose when to trigger a rebalance using the kafka-leader-election.sh command line tool.

Note
The Grafana dashboards provided with Strimzi show metrics for under-replicated partitions and partitions that do not have an active leader.
Unclean leader election

Leader election to an in-sync replica is considered clean because it guarantees no loss of data. And this is what happens by default. But what if there is no in-sync replica to take on leadership? Perhaps the ISR (in-sync replica) only contained the leader when the leader’s disk died. If a minimum number of in-sync replicas is not set, and there are no followers in sync with the partition leader when its hard drive fails irrevocably, data is already lost. Not only that, but a new leader cannot be elected because there are no in-sync followers.

You can configure how Kafka handles leader failure:

# ...
unclean.leader.election.enable=false
# ...

Unclean leader election is disabled by default, which means that out-of-sync replicas cannot become leaders. With clean leader election, if no other broker was in the ISR when the old leader was lost, Kafka waits until that leader is back online before messages can be written or read. Unclean leader election means out-of-sync replicas can become leaders, but you risk losing messages. The choice you make depends on whether your requirements favor availability or durability.

You can override the default configuration for specific topics at the topic level. If you cannot afford the risk of data loss, then leave the default configuration.

Avoiding unnecessary consumer group rebalances

For consumers joining a new consumer group, you can add a delay so that unnecessary rebalances to the broker are avoided:

# ...
group.initial.rebalance.delay.ms=3000
# ...

The delay is the amount of time that the coordinator waits for members to join. The longer the delay, the more likely it is that all the members will join in time and avoid a rebalance. But the delay also prevents the group from consuming until the period has ended.

10.6.2. Kafka producer configuration tuning

Use a basic producer configuration with optional properties that are tailored to specific use cases.

Adjusting your configuration to maximize throughput might increase latency or vice versa. You will need to experiment and tune your producer configuration to get the balance you need.

Basic producer configuration

Connection and serializer properties are required for every producer. Generally, it is good practice to add a client id for tracking, and use compression on the producer to reduce batch sizes in requests.

In a basic producer configuration:

  • The order of messages in a partition is not guaranteed.

  • The acknowledgment of messages reaching the broker does not guarantee durability.

Basic producer configuration properties
# ...
bootstrap.servers=localhost:9092 (1)
key.serializer=org.apache.kafka.common.serialization.StringSerializer (2)
value.serializer=org.apache.kafka.common.serialization.StringSerializer (3)
client.id=my-client (4)
compression.type=gzip (5)
# ...
  1. (Required) Tells the producer to connect to a Kafka cluster using a host:port bootstrap server address for a Kafka broker. The producer uses the address to discover and connect to all brokers in the cluster. Use a comma-separated list to specify two or three addresses in case a server is down, but it’s not necessary to provide a list of all the brokers in the cluster.

  2. (Required) Serializer to transform the key of each message to bytes prior to them being sent to a broker.

  3. (Required) Serializer to transform the value of each message to bytes prior to them being sent to a broker.

  4. (Optional) The logical name for the client, which is used in logs and metrics to identify the source of a request.

  5. (Optional) The codec for compressing messages, which are sent and might be stored in compressed format and then decompressed when reaching a consumer. Compression is useful for improving throughput and reducing the load on storage, but might not be suitable for low latency applications where the cost of compression or decompression could be prohibitive.

Data durability

You can apply greater data durability, to minimize the likelihood that messages are lost, using message delivery acknowledgments.

# ...
acks=all (1)
# ...
  1. Specifying acks=all forces a partition leader to replicate messages to a certain number of followers before acknowledging that the message request was successfully received. Because of the additional checks, acks=all increases the latency between the producer sending a message and receiving acknowledgment.

The number of brokers which need to have appended the messages to their logs before the acknowledgment is sent to the producer is determined by the topic’s min.insync.replicas configuration. A typical starting point is to have a topic replication factor of 3, with two in-sync replicas on other brokers. In this configuration, the producer can continue unaffected if a single broker is unavailable. If a second broker becomes unavailable, the producer won’t receive acknowledgments and won’t be able to produce more messages.

Topic configuration to support acks=all
# ...
min.insync.replicas=2 (1)
# ...
  1. Use 2 in-sync replicas. The default is 1.

Note
If the system fails, there is a risk of unsent data in the buffer being lost.
Ordered delivery

Idempotent producers avoid duplicates as messages are delivered exactly once. IDs and sequence numbers are assigned to messages to ensure the order of delivery, even in the event of failure. If you are using acks=all for data consistency, enabling idempotency makes sense for ordered delivery.

Ordered delivery with idempotency
# ...
enable.idempotence=true (1)
max.in.flight.requests.per.connection=5 (2)
acks=all (3)
retries=2147483647 (4)
# ...
  1. Set to true to enable the idempotent producer.

  2. With idempotent delivery the number of in-flight requests may be greater than 1 while still providing the message ordering guarantee. The default is 5 in-flight requests.

  3. Set acks to all.

  4. Set the number of attempts to resend a failed message request.

If you are not using acks=all and idempotency because of the performance cost, set the number of in-flight (unacknowledged) requests to 1 to preserve ordering. Otherwise, a situation is possible where Message-A fails only to succeed after Message-B was already written to the broker.

Ordered delivery without idempotency
# ...
enable.idempotence=false (1)
max.in.flight.requests.per.connection=1 (2)
retries=2147483647
# ...
  1. Set to false to disable the idempotent producer.

  2. Set the number of in-flight requests to exactly 1.

Reliability guarantees

Idempotence is useful for exactly once writes to a single partition. Transactions, when used with idempotence, allow exactly once writes across multiple partitions.

Transactions guarantee that messages using the same transactional ID are produced once, and either all are successfully written to the respective logs or none of them are.

# ...
enable.idempotence=true
max.in.flight.requests.per.connection=5
acks=all
retries=2147483647
transactional.id=UNIQUE-ID (1)
transaction.timeout.ms=900000 (2)
# ...
  1. Specify a unique transactional ID.

  2. Set the maximum allowed time for transactions in milliseconds before a timeout error is returned. The default is 900000 or 15 minutes.

The choice of transactional.id is important in order that the transactional guarantee is maintained. Each transactional id should be used for a unique set of topic partitions. For example, this can be achieved using an external mapping of topic partition names to transactional ids, or by computing the transactional id from the topic partition names using a function that avoids collisions.

Optimizing throughput and latency

Usually, the requirement of a system is to satisfy a particular throughput target for a proportion of messages within a given latency. For example, targeting 500,000 messages per second with 95% of messages being acknowledged within 2 seconds.

It’s likely that the messaging semantics (message ordering and durability) of your producer are defined by the requirements for your application. For instance, it’s possible that you don’t have the option of using acks=0 or acks=1 without breaking some important property or guarantee provided by your application.

Broker restarts have a significant impact on high percentile statistics. For example, over a long period the 99th percentile latency is dominated by behavior around broker restarts. This is worth considering when designing benchmarks or comparing performance numbers from benchmarking with performance numbers seen in production.

Depending on your objective, Kafka offers a number of configuration parameters and techniques for tuning producer performance for throughput and latency.

Message batching (linger.ms and batch.size)

Message batching delays sending messages in the hope that more messages destined for the same broker will be sent, allowing them to be batched into a single produce request. Batching is a compromise between higher latency in return for higher throughput. Time-based batching is configured using linger.ms, and size-based batching is configured using batch.size.

Compression (compression.type)

Message compression adds latency in the producer (CPU time spent compressing the messages), but makes requests (and potentially disk writes) smaller, which can increase throughput. Whether compression is worthwhile, and the best compression to use, will depend on the messages being sent. Compression happens on the thread which calls KafkaProducer.send(), so if the latency of this method matters for your application you should consider using more threads.

Pipelining (max.in.flight.requests.per.connection)

Pipelining means sending more requests before the response to a previous request has been received. In general more pipelining means better throughput, up to a threshold at which other effects, such as worse batching, start to counteract the effect on throughput.

Lowering latency

When your application calls KafkaProducer.send() the messages are:

  • Processed by any interceptors

  • Serialized

  • Assigned to a partition

  • Compressed

  • Added to a batch of messages in a per-partition queue

At which point the send() method returns. So the time send() is blocked is determined by:

  • The time spent in the interceptors, serializers and partitioner

  • The compression algorithm used

  • The time spent waiting for a buffer to use for compression

Batches will remain in the queue until one of the following occurs:

  • The batch is full (according to batch.size)

  • The delay introduced by linger.ms has passed

  • The sender is about to send message batches for other partitions to the same broker, and it is possible to add this batch too

  • The producer is being flushed or closed

Look at the configuration for batching and buffering to mitigate the impact of send() blocking on latency.

# ...
linger.ms=100 (1)
batch.size=16384 (2)
buffer.memory=33554432 (3)
# ...
  1. The linger property adds a delay in milliseconds so that larger batches of messages are accumulated and sent in a request. The default is 0'.

  2. If a maximum batch.size in bytes is used, a request is sent when the maximum is reached, or messages have been queued for longer than linger.ms (whichever comes sooner). Adding the delay allows batches to accumulate messages up to the batch size.

  3. The buffer size must be at least as big as the batch size, and be able to accommodate buffering, compression and in-flight requests.

Increasing throughput

Improve throughput of your message requests by adjusting the maximum time to wait before a message is delivered and completes a send request.

You can also direct messages to a specified partition by writing a custom partitioner to replace the default.

# ...
delivery.timeout.ms=120000 (1)
partitioner.class=my-custom-partitioner (2)

# ...
  1. The maximum time in milliseconds to wait for a complete send request. You can set the value to MAX_LONG to delegate to Kafka an indefinite number of retries. The default is 120000 or 2 minutes.

  2. Specify the class name of the custom partitioner.

10.6.3. Kafka consumer configuration tuning

Use a basic consumer configuration with optional properties that are tailored to specific use cases.

When tuning your consumers your primary concern will be ensuring that they cope efficiently with the amount of data ingested. As with the producer tuning, be prepared to make incremental changes until the consumers operate as expected.

Basic consumer configuration

Connection and deserializer properties are required for every consumer. Generally, it is good practice to add a client id for tracking.

In a consumer configuration, irrespective of any subsequent configuration:

  • The consumer fetches from a given offset and consumes the messages in order, unless the offset is changed to skip or re-read messages.

  • The broker does not know if the consumer processed the responses, even when committing offsets to Kafka, because the offsets might be sent to a different broker in the cluster.

Basic consumer configuration properties
# ...
bootstrap.servers=localhost:9092 (1)
key.deserializer=org.apache.kafka.common.serialization.StringDeserializer  (2)
value.deserializer=org.apache.kafka.common.serialization.StringDeserializer  (3)
client.id=my-client (4)
group.id=my-group-id (5)
# ...
  1. (Required) Tells the consumer to connect to a Kafka cluster using a host:port bootstrap server address for a Kafka broker. The consumer uses the address to discover and connect to all brokers in the cluster. Use a comma-separated list to specify two or three addresses in case a server is down, but it is not necessary to provide a list of all the brokers in the cluster. If you are using a loadbalancer service to expose the Kafka cluster, you only need the address for the service because the availability is handled by the loadbalancer.

  2. (Required) Deserializer to transform the bytes fetched from the Kafka broker into message keys.

  3. (Required) Deserializer to transform the bytes fetched from the Kafka broker into message values.

  4. (Optional) The logical name for the client, which is used in logs and metrics to identify the source of a request. The id can also be used to throttle consumers based on processing time quotas.

  5. (Conditional) A group id is required for a consumer to be able to join a consumer group.

Scaling data consumption using consumer groups

Consumer groups share a typically large data stream generated by one or multiple producers from a given topic. Consumers are grouped using a group.id property, allowing messages to be spread across the members. One of the consumers in the group is elected leader and decides how the partitions are assigned to the consumers in the group. Each partition can only be assigned to a single consumer.

If you do not already have as many consumers as partitions, you can scale data consumption by adding more consumer instances with the same group.id. Adding more consumers to a group than there are partitions will not help throughput, but it does mean that there are consumers on standby should one stop functioning. If you can meet throughput goals with fewer consumers, you save on resources.

Consumers within the same consumer group send offset commits and heartbeats to the same broker. So the greater the number of consumers in the group, the higher the request load on the broker.

# ...
group.id=my-group-id (1)
# ...
  1. Add a consumer to a consumer group using a group id.

Message ordering guarantees

Kafka brokers receive fetch requests from consumers that ask the broker to send messages from a list of topics, partitions and offset positions.

A consumer observes messages in a single partition in the same order that they were committed to the broker, which means that Kafka only provides ordering guarantees for messages in a single partition. Conversely, if a consumer is consuming messages from multiple partitions, the order of messages in different partitions as observed by the consumer does not necessarily reflect the order in which they were sent.

If you want a strict ordering of messages from one topic, use one partition per consumer.

Optimizing throughput and latency

Control the number of messages returned when your client application calls KafkaConsumer.poll().

Use the fetch.max.wait.ms and fetch.min.bytes properties to increase the minimum amount of data fetched by the consumer from the Kafka broker. Time-based batching is configured using fetch.max.wait.ms, and size-based batching is configured using fetch.min.bytes.

If CPU utilization in the consumer or broker is high, it might be because there are too many requests from the consumer. You can adjust fetch.max.wait.ms and fetch.min.bytes properties higher so that there are fewer requests and messages are delivered in bigger batches. By adjusting higher, throughput is improved with some cost to latency. You can also adjust higher if the amount of data being produced is low.

For example, if you set fetch.max.wait.ms to 500ms and fetch.min.bytes to 16384 bytes, when Kafka receives a fetch request from the consumer it will respond when the first of either threshold is reached.

Conversely, you can adjust the fetch.max.wait.ms and fetch.min.bytes properties lower to improve end-to-end latency.

# ...
fetch.max.wait.ms=500 (1)
fetch.min.bytes=16384 (2)
# ...
  1. The maximum time in milliseconds the broker will wait before completing fetch requests. The default is 500 milliseconds.

  2. If a minimum batch size in bytes is used, a request is sent when the minimum is reached, or messages have been queued for longer than fetch.max.wait.ms (whichever comes sooner). Adding the delay allows batches to accumulate messages up to the batch size.

Lowering latency by increasing the fetch request size

Use the fetch.max.bytes and max.partition.fetch.bytes properties to increase the maximum amount of data fetched by the consumer from the Kafka broker.

The fetch.max.bytes property sets a maximum limit in bytes on the amount of data fetched from the broker at one time.

The max.partition.fetch.bytes sets a maximum limit in bytes on how much data is returned for each partition, which must always be larger than the number of bytes set in the broker or topic configuration for max.message.bytes.

The maximum amount of memory a client can consume is calculated approximately as:

NUMBER-OF-BROKERS * fetch.max.bytes and NUMBER-OF-PARTITIONS * max.partition.fetch.bytes

If memory usage can accommodate it, you can increase the values of these two properties. By allowing more data in each request, latency is improved as there are fewer fetch requests.

# ...
fetch.max.bytes=52428800 (1)
max.partition.fetch.bytes=1048576 (2)
# ...
  1. The maximum amount of data in bytes returned for a fetch request.

  2. The maximum amount of data in bytes returned for each partition.

Avoiding data loss or duplication when committing offsets

The Kafka auto-commit mechanism allows a consumer to commit the offsets of messages automatically. If enabled, the consumer will commit offsets received from polling the broker at 5000ms intervals.

The auto-commit mechanism is convenient, but it introduces a risk of data loss and duplication. If a consumer has fetched and transformed a number of messages, but the system crashes with processed messages in the consumer buffer when performing an auto-commit, that data is lost. If the system crashes after processing the messages, but before performing the auto-commit, the data is duplicated on another consumer instance after rebalancing.

Auto-committing can avoid data loss only when all messages are processed before the next poll to the broker, or the consumer closes.

To minimize the likelihood of data loss or duplication, you can set enable.auto.commit to false and develop your client application to have more control over committing offsets. Or you can use auto.commit.interval.ms to decrease the intervals between commits.

# ...
enable.auto.commit=false (1)
# ...
  1. Auto commit is set to false to provide more control over committing offsets.

By setting to enable.auto.commit to false, you can commit offsets after all processing has been performed and the message has been consumed. For example, you can set up your application to call the Kafka commitSync and commitAsync commit APIs.

The commitSync API commits the offsets in a message batch returned from polling. You call the API when you are finished processing all the messages in the batch. If you use the commitSync API, the application will not poll for new messages until the last offset in the batch is committed. If this negatively affects throughput, you can commit less frequently, or you can use the commitAsync API. The commitAsync API does not wait for the broker to respond to a commit request, but risks creating more duplicates when rebalancing. A common approach is to combine both commit APIs in an application, with the commitSync API used just before shutting the consumer down or rebalancing to make sure the final commit is successful.

Controlling transactional messages

Consider using transactional ids and enabling idempotence (enable.idempotence=true) on the producer side to guarantee exactly-once delivery. On the consumer side, you can then use the isolation.level property to control how transactional messages are read by the consumer.

The isolation.level property has two valid values:

  • read_committed

  • read_uncommitted (default)

Use read_committed to ensure that only transactional messages that have been committed are read by the consumer. However, this will cause an increase in end-to-end latency, because the consumer will not be able to return a message until the brokers have written the transaction markers that record the result of the transaction (committed or aborted).

# ...
enable.auto.commit=false
isolation.level=read_committed (1)
# ...
  1. Set to read_committed so that only committed messages are read by the consumer.

Recovering from failure to avoid data loss

Use the session.timeout.ms and heartbeat.interval.ms properties to configure the time taken to check and recover from consumer failure within a consumer group.

The session.timeout.ms property specifies the maximum amount of time in milliseconds a consumer within a consumer group can be out of contact with a broker before being considered inactive and a rebalancing is triggered between the active consumers in the group. When the group rebalances, the partitions are reassigned to the members of the group.

The heartbeat.interval.ms property specifies the interval in milliseconds between heartbeat checks to the consumer group coordinator to indicate that the consumer is active and connected. The heartbeat interval must be lower, usually by a third, than the session timeout interval.

If you set the session.timeout.ms property lower, failing consumers are detected earlier, and rebalancing can take place quicker. However, take care not to set the timeout so low that the broker fails to receive a heartbeat in time and triggers an unnecessary rebalance.

Decreasing the heartbeat interval reduces the chance of accidental rebalancing, but more frequent heartbeats increases the overhead on broker resources.

Managing offset policy

Use the auto.offset.reset property to control how a consumer behaves when no offsets have been committed, or a committed offset is no longer valid or deleted.

Suppose you deploy a consumer application for the first time, and it reads messages from an existing topic. Because this is the first time the group.id is used, the __consumer_offsets topic does not contain any offset information for this application. The new application can start processing all existing messages from the start of the log or only new messages. The default reset value is latest, which starts at the end of the partition, and consequently means some messages are missed. To avoid data loss, but increase the amount of processing, set auto.offset.reset to earliest to start at the beginning of the partition.

Also consider using the earliest option to avoid messages being lost when the offsets retention period (offsets.retention.minutes) configured for a broker has ended. If a consumer group or standalone consumer is inactive and commits no offsets during the retention period, previously committed offsets are deleted from __consumer_offsets.

# ...
heartbeat.interval.ms=3000 (1)
session.timeout.ms=10000 (2)
auto.offset.reset=earliest (3)
# ...
  1. Adjust the heartbeat interval lower according to anticipated rebalances.

  2. If no heartbeats are received by the Kafka broker before the timeout duration expires, the consumer is removed from the consumer group and a rebalance is initiated. If the broker configuration has a group.min.session.timeout.ms and group.max.session.timeout.ms, the session timeout value must be within that range.

  3. Set to earliest to return to the start of a partition and avoid data loss if offsets were not committed.

If the amount of data returned in a single fetch request is large, a timeout might occur before the consumer has processed it. In this case, you can lower max.partition.fetch.bytes or increase session.timeout.ms.

Minimizing the impact of rebalances

The rebalancing of a partition between active consumers in a group is the time it takes for:

  • Consumers to commit their offsets

  • The new consumer group to be formed

  • The group leader to assign partitions to group members

  • The consumers in the group to receive their assignments and start fetching

Clearly, the process increases the downtime of a service, particularly when it happens repeatedly during a rolling restart of a consumer group cluster.

In this situation, you can use the concept of static membership to reduce the number of rebalances. Rebalancing assigns topic partitions evenly among consumer group members. Static membership uses persistence so that a consumer instance is recognized during a restart after a session timeout.

The consumer group coordinator can identify a new consumer instance using a unique id that is specified using the group.instance.id property. During a restart, the consumer is assigned a new member id, but as a static member it continues with the same instance id, and the same assignment of topic partitions is made.

If the consumer application does not make a call to poll at least every max.poll.interval.ms milliseconds, the consumer is considered to be failed, causing a rebalance. If the application cannot process all the records returned from poll in time, you can avoid a rebalance by using the max.poll.interval.ms property to specify the interval in milliseconds between polls for new messages from a consumer. Or you can use the max.poll.records property to set a maximum limit on the number of records returned from the consumer buffer, allowing your application to process fewer records within the max.poll.interval.ms limit.

# ...
group.instance.id=_UNIQUE-ID_ (1)
max.poll.interval.ms=300000 (2)
max.poll.records=500 (3)
# ...
  1. The unique instance id ensures that a new consumer instance receives the same assignment of topic partitions.

  2. Set the interval to check the consumer is continuing to process messages.

  3. Sets the number of processed records returned from the consumer.

10.7. Uninstalling Strimzi

This procedure describes how to uninstall Strimzi and remove resources related to the deployment.

Prerequisites

In order to perform this procedure, identify resources created specifically for a deployment and referenced from the Strimzi resource.

Such resources include:

  • Secrets (Custom CAs and certificates, Kafka Connect secrets, and other Kafka secrets)

  • Logging ConfigMaps (of type external)

These are resources referenced by Kafka, KafkaConnect, KafkaConnectS2I, KafkaMirrorMaker, or KafkaBridge configuration.

Procedure
  1. Delete the Cluster Operator Deployment, related CustomResourceDefinitions, and RBAC resources:

    kubectl delete -f install/cluster-operator
    Warning
    Deleting CustomResourceDefinitions results in the garbage collection of the corresponding custom resources (Kafka, KafkaConnect, KafkaConnectS2I, KafkaMirrorMaker, or KafkaBridge) and the resources dependent on them (Deployments, StatefulSets, and other dependent resources).
  2. Delete the resources you identified in the prerequisites.

10.8. Frequently asked questions

Why do I need cluster administrator privileges to install Strimzi?

To install Strimzi, you need to be able to create the following cluster-scoped resources:

  • Custom Resource Definitions (CRDs) to instruct Kubernetes about resources that are specific to Strimzi, such as Kafka and KafkaConnect

  • ClusterRoles and ClusterRoleBindings

Cluster-scoped resources, which are not scoped to a particular Kubernetes namespace, typically require cluster administrator privileges to install.

As a cluster administrator, you can inspect all the resources being installed (in the /install/ directory) to ensure that the ClusterRoles do not grant unnecessary privileges.

After installation, the Cluster Operator runs as a regular Deployment, so any standard (non-admin) Kubernetes user with privileges to access the Deployment can configure it. The cluster administrator can grant standard users the privileges necessary to manage Kafka custom resources.

See also:

Why does the Cluster Operator need to create ClusterRoleBindings?

Kubernetes has built-in privilege escalation prevention, which means that the Cluster Operator cannot grant privileges it does not have itself, specifically, it cannot grant such privileges in a namespace it cannot access. Therefore, the Cluster Operator must have the privileges necessary for all the components it orchestrates.

The Cluster Operator needs to be able to grant access so that:

  • The Topic Operator can manage KafkaTopics, by creating Roles and RoleBindings in the namespace that the operator runs in

  • The User Operator can manage KafkaUsers, by creating Roles and RoleBindings in the namespace that the operator runs in

  • The failure domain of a Node is discovered by Strimzi, by creating a ClusterRoleBinding

When using rack-aware partition assignment, the broker pod needs to be able to get information about the Node it is running on, for example, the Availability Zone in Amazon AWS. A Node is a cluster-scoped resource, so access to it can only be granted through a ClusterRoleBinding, not a namespace-scoped RoleBinding.

Can standard Kubernetes users create Kafka custom resources?

By default, standard Kubernetes users will not have the privileges necessary to manage the custom resources handled by the Cluster Operator. The cluster administrator can grant a user the necessary privileges using Kubernetes RBAC resources.

For more information, see Designating Strimzi administrators in the Deploying and Upgrading Strimzi guide.

What do the failed to acquire lock warnings in the log mean?

For each cluster, the Cluster Operator executes only one operation at a time. The Cluster Operator uses locks to make sure that there are never two parallel operations running for the same cluster. Other operations must wait until the current operation completes before the lock is released.

INFO

Examples of cluster operations include cluster creation, rolling update, scale down , and scale up.

If the waiting time for the lock takes too long, the operation times out and the following warning message is printed to the log:

2018-03-04 17:09:24 WARNING AbstractClusterOperations:290 - Failed to acquire lock for kafka cluster lock::kafka::myproject::my-cluster

Depending on the exact configuration of STRIMZI_FULL_RECONCILIATION_INTERVAL_MS and STRIMZI_OPERATION_TIMEOUT_MS, this warning message might appear occasionally without indicating any underlying issues. Operations that time out are picked up in the next periodic reconciliation, so that the operation can acquire the lock and execute again.

Should this message appear periodically, even in situations when there should be no other operations running for a given cluster, it might indicate that the lock was not properly released due to an error. If this is the case, try restarting the Cluster Operator.

Why is hostname verification failing when connecting to NodePorts using TLS?

Currently, off-cluster access using NodePorts with TLS encryption enabled does not support TLS hostname verification. As a result, the clients that verify the hostname will fail to connect. For example, the Java client will fail with the following exception:

Caused by: java.security.cert.CertificateException: No subject alternative names matching IP address 168.72.15.231 found
 at sun.security.util.HostnameChecker.matchIP(HostnameChecker.java:168)
 at sun.security.util.HostnameChecker.match(HostnameChecker.java:94)
 at sun.security.ssl.X509TrustManagerImpl.checkIdentity(X509TrustManagerImpl.java:455)
 at sun.security.ssl.X509TrustManagerImpl.checkIdentity(X509TrustManagerImpl.java:436)
 at sun.security.ssl.X509TrustManagerImpl.checkTrusted(X509TrustManagerImpl.java:252)
 at sun.security.ssl.X509TrustManagerImpl.checkServerTrusted(X509TrustManagerImpl.java:136)
 at sun.security.ssl.ClientHandshaker.serverCertificate(ClientHandshaker.java:1501)
 ... 17 more

To connect, you must disable hostname verification. In the Java client, you can do this by setting the configuration option ssl.endpoint.identification.algorithm to an empty string.

When configuring the client using a properties file, you can do it this way:

ssl.endpoint.identification.algorithm=

When configuring the client directly in Java, set the configuration option to an empty string:

props.put("ssl.endpoint.identification.algorithm", "");

11. Custom resource API reference

11.1. Common configuration properties

Common configuration properties apply to more than one resource.

11.1.1. replicas

Use the replicas property to configure replicas.

The type of replication depends on the resource.

  • KafkaTopic uses a replication factor to configure the number of replicas of each partition within a Kafka cluster.

  • Kafka components use replicas to configure the number of pods in a deployment to provide better availability and scalability.

Note
When running a Kafka component on Kubernetes it may not be necessary to run multiple replicas for high availability. When the node where the component is deployed crashes, Kubernetes will automatically reschedule the Kafka component pod to a different node. However, running Kafka components with multiple replicas can provide faster failover times as the other nodes will be up and running.

11.1.2. bootstrapServers

Use the bootstrapServers property to configure a list of bootstrap servers.

The bootstrap server lists can refer to Kafka clusters that are not deployed in the same Kubernetes cluster. They can also refer to a Kafka cluster not deployed by Strimzi.

If on the same Kubernetes cluster, each list must ideally contain the Kafka cluster bootstrap service which is named CLUSTER-NAME-kafka-bootstrap and a port number. If deployed by Strimzi but on different Kubernetes clusters, the list content depends on the approach used for exposing the clusters (routes, ingress, nodeports or loadbalancers).

When using Kafka with a Kafka cluster not managed by Strimzi, you can specify the bootstrap servers list according to the configuration of the given cluster.

11.1.3. ssl

Use the three allowed ssl configuration options for client connection using a specific cipher suite for a TLS version. A cipher suite combines algorithms for secure connection and data transfer.

You can also configure the ssl.endpoint.identification.algorithm property to enable or disable hostname verification.

Example SSL configuration
# ...
spec:
  config:
    ssl.cipher.suites: "TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384" (1)
    ssl.enabled.protocols: "TLSv1.2" (2)
    ssl.protocol: "TLSv1.2" (3)
    ssl.endpoint.identification.algorithm: HTTPS (4)
# ...
  1. The cipher suite for TLS using a combination of ECDHE key exchange mechanism, RSA authentication algorithm, AES bulk encyption algorithm and SHA384 MAC algorithm.

  2. The SSl protocol TLSv1.2 is enabled.

  3. Specifies the TLSv1.2 protocol to generate the SSL context. Allowed values are TLSv1.1 and TLSv1.2.

  4. Hostname verification is enabled by setting to HTTPS. An empty string disables the verification.

11.1.4. trustedCertificates

Having set tls to configure TLS encryption, use the trustedCertificates property to provide a list of secrets with key names under which the certificates are stored in X.509 format.

You can use the secrets created by the Cluster Operator for the Kafka cluster, or you can create your own TLS certificate file, then create a Secret from the file:

kubectl create secret generic MY-SECRET \
--from-file=MY-TLS-CERTIFICATE-FILE.crt
Example TLS encryption configuration
tls:
  trustedCertificates:
    - secretName: my-cluster-cluster-cert
      certificate: ca.crt
    - secretName: my-cluster-cluster-cert
      certificate: ca2.crt

If certificates are stored in the same secret, it can be listed multiple times.

If you want to enable TLS, but use the default set of public certification authorities shipped with Java, you can specify trustedCertificates as an empty array:

Example of enabling TLS with the default Java certificates
tls:
  trustedCertificates: []

For information on configuring TLS client authentication, see KafkaClientAuthenticationTls schema reference.

11.1.5. resources

You request CPU and memory resources for components. Limits specify the maximum resources that can be consumed by a given container.

Resource requests and limits for the Topic Operator and User Operator are set in the Kafka resource.

Use the reources.requests and resources.limits properties to configure resource requests and limits.

For every deployed container, Strimzi allows you to request specific resources and define the maximum consumption of those resources.

Strimzi supports requests and limits for the following types of resources:

  • cpu

  • memory

Strimzi uses the Kubernetes syntax for specifying these resources.

For more information about managing computing resources on Kubernetes, see Managing Compute Resources for Containers.

Resource requests

Requests specify the resources to reserve for a given container. Reserving the resources ensures that they are always available.

Important
If the resource request is for more than the available free resources in the Kubernetes cluster, the pod is not scheduled.

A request may be configured for one or more supported resources.

Example resource requests configuration
# ...
resources:
  requests:
    cpu: 12
    memory: 64Gi
# ...

Resource limits

Limits specify the maximum resources that can be consumed by a given container. The limit is not reserved and might not always be available. A container can use the resources up to the limit only when they are available. Resource limits should be always higher than the resource requests.

A resource may be configured for one or more supported limits.

Example resource limits configuration
# ...
resources:
  limits:
    cpu: 12
    memory: 64Gi
# ...

Supported CPU formats

CPU requests and limits are supported in the following formats:

  • Number of CPU cores as integer (5 CPU core) or decimal (2.5 CPU core).

  • Number or millicpus / millicores (100m) where 1000 millicores is the same 1 CPU core.

Example CPU units
# ...
resources:
  requests:
    cpu: 500m
  limits:
    cpu: 2.5
# ...
Note
The computing power of 1 CPU core may differ depending on the platform where Kubernetes is deployed.

For more information on CPU specification, see the Meaning of CPU.

Supported memory formats

Memory requests and limits are specified in megabytes, gigabytes, mebibytes, and gibibytes.

  • To specify memory in megabytes, use the M suffix. For example 1000M.

  • To specify memory in gigabytes, use the G suffix. For example 1G.

  • To specify memory in mebibytes, use the Mi suffix. For example 1000Mi.

  • To specify memory in gibibytes, use the Gi suffix. For example 1Gi.

Example resources using different memory units
# ...
resources:
  requests:
    memory: 512Mi
  limits:
    memory: 2Gi
# ...

For more details about memory specification and additional supported units, see Meaning of memory.

11.1.6. image

Use the image property to configure the container image used by the component.

Overriding container images is recommended only in special situations where you need to use a different container registry or a customized image.

For example, if your network does not allow access to the container repository used by Strimzi, you can copy the Strimzi images or build them from the source. However, if the configured image is not compatible with Strimzi images, it might not work properly.

A copy of the container image might also be customized and used for debugging.

You can specify which container image to use for a component using the image property in the following resources:

  • Kafka.spec.kafka

  • Kafka.spec.zookeeper

  • Kafka.spec.entityOperator.topicOperator

  • Kafka.spec.entityOperator.userOperator

  • Kafka.spec.entityOperator.tlsSidecar

  • Kafka.spec.jmxTrans

  • KafkaConnect.spec

  • KafkaConnectS2I.spec

  • KafkaMirrorMaker.spec

  • KafkaMirrorMaker2.spec

  • KafkaBridge.spec

Configuring the image property for Kafka, Kafka Connect, and Kafka MirrorMaker

Kafka, Kafka Connect (including Kafka Connect with S2I support), and Kafka MirrorMaker support multiple versions of Kafka. Each component requires its own image. The default images for the different Kafka versions are configured in the following environment variables:

  • STRIMZI_KAFKA_IMAGES

  • STRIMZI_KAFKA_CONNECT_IMAGES

  • STRIMZI_KAFKA_CONNECT_S2I_IMAGES

  • STRIMZI_KAFKA_MIRROR_MAKER_IMAGES

These environment variables contain mappings between the Kafka versions and their corresponding images. The mappings are used together with the image and version properties:

  • If neither image nor version are given in the custom resource then the version will default to the Cluster Operator’s default Kafka version, and the image will be the one corresponding to this version in the environment variable.

  • If image is given but version is not, then the given image is used and the version is assumed to be the Cluster Operator’s default Kafka version.

  • If version is given but image is not, then the image that corresponds to the given version in the environment variable is used.

  • If both version and image are given, then the given image is used. The image is assumed to contain a Kafka image with the given version.

The image and version for the different components can be configured in the following properties:

  • For Kafka in spec.kafka.image and spec.kafka.version.

  • For Kafka Connect, Kafka Connect S2I, and Kafka MirrorMaker in spec.image and spec.version.

Warning
It is recommended to provide only the version and leave the image property unspecified. This reduces the chance of making a mistake when configuring the custom resource. If you need to change the images used for different versions of Kafka, it is preferable to configure the Cluster Operator’s environment variables.

Configuring the image property in other resources

For the image property in the other custom resources, the given value will be used during deployment. If the image property is missing, the image specified in the Cluster Operator configuration will be used. If the image name is not defined in the Cluster Operator configuration, then the default value will be used.

  • For Topic Operator:

    1. Container image specified in the STRIMZI_DEFAULT_TOPIC_OPERATOR_IMAGE environment variable from the Cluster Operator configuration.

    2. quay.io/strimzi/operator:0.24.0 container image.

  • For User Operator:

    1. Container image specified in the STRIMZI_DEFAULT_USER_OPERATOR_IMAGE environment variable from the Cluster Operator configuration.

    2. quay.io/strimzi/operator:0.24.0 container image.

  • For Entity Operator TLS sidecar:

    1. Container image specified in the STRIMZI_DEFAULT_TLS_SIDECAR_ENTITY_OPERATOR_IMAGE environment variable from the Cluster Operator configuration.

    2. quay.io/strimzi/kafka:0.24.0-kafka-2.8.0 container image.

  • For Kafka Exporter:

    1. Container image specified in the STRIMZI_DEFAULT_KAFKA_EXPORTER_IMAGE environment variable from the Cluster Operator configuration.

    2. quay.io/strimzi/kafka:0.24.0-kafka-2.8.0 container image.

  • For Kafka Bridge:

    1. Container image specified in the STRIMZI_DEFAULT_KAFKA_BRIDGE_IMAGE environment variable from the Cluster Operator configuration.

    2. quay.io/strimzi/kafka-bridge:0.20.1 container image.

  • For Kafka broker initializer:

    1. Container image specified in the STRIMZI_DEFAULT_KAFKA_INIT_IMAGE environment variable from the Cluster Operator configuration.

    2. quay.io/strimzi/operator:0.24.0 container image.

  • For Kafka jmxTrans:

    1. Container image specified in the STRIMZI_DEFAULT_JMXTRANS_IMAGE environment variable from the Cluster Operator configuration.

    2. quay.io/strimzi/jmxtrans:0.24.0 container image.

Example of container image configuration
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
    image: my-org/my-image:latest
    # ...
  zookeeper:
    # ...

11.1.7. livenessProbe and readinessProbe healthchecks

Use the livenessProbe and readinessProbe properties to configure healthcheck probes supported in Strimzi.

Healthchecks are periodical tests which verify the health of an application. When a Healthcheck probe fails, Kubernetes assumes that the application is not healthy and attempts to fix it.

For more details about the probes, see Configure Liveness and Readiness Probes.

Both livenessProbe and readinessProbe support the following options:

  • initialDelaySeconds

  • timeoutSeconds

  • periodSeconds

  • successThreshold

  • failureThreshold

Example of liveness and readiness probe configuration
# ...
readinessProbe:
  initialDelaySeconds: 15
  timeoutSeconds: 5
livenessProbe:
  initialDelaySeconds: 15
  timeoutSeconds: 5
# ...

For more information about the livenessProbe and readinessProbe options, see Probe schema reference.

11.1.8. metricsConfig

Use the metricsConfig property to enable and configure Prometheus metrics.

The metricsConfig property contains a reference to a ConfigMap containing additional configuration for the Prometheus JMX exporter. Strimzi supports Prometheus metrics using Prometheus JMX exporter to convert the JMX metrics supported by Apache Kafka and ZooKeeper to Prometheus metrics.

To enable Prometheus metrics export without further configuration, you can reference a ConfigMap containing an empty file under metricsConfig.valueFrom.configMapKeyRef.key. When referencing an empty file, all metrics are exposed as long as they have not been renamed.

Example ConfigMap with metrics configuration for Kafka
kind: ConfigMap
apiVersion: v1
metadata:
  name: my-configmap
data:
  my-key: |
    lowercaseOutputName: true
    rules:
    # Special cases and very specific rules
    - pattern: kafka.server<type=(.+), name=(.+), clientId=(.+), topic=(.+), partition=(.*)><>Value
      name: kafka_server_$1_$2
      type: GAUGE
      labels:
       clientId: "$3"
       topic: "$4"
       partition: "$5"
    # further configuration
Example metrics configuration for Kafka
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
    metricsConfig:
      type: jmxPrometheusExporter
      valueFrom:
        configMapKeyRef:
          name: my-config-map
          key: my-key
    # ...
  zookeeper:
    # ...

When metrics are enabled, they are exposed on port 9404.

When the metricsConfig (or deprecated metrics) property is not defined in the resource, the Prometheus metrics are disabled.

For more information about setting up and deploying Prometheus and Grafana, see Introducing Metrics to Kafka in the Deploying and Upgrading Strimzi guide.

11.1.9. jvmOptions

The following Strimzi components run inside a Java Virtual Machine (JVM):

  • Apache Kafka

  • Apache ZooKeeper

  • Apache Kafka Connect

  • Apache Kafka MirrorMaker

  • Strimzi Kafka Bridge

To optimize their performance on different platforms and architectures, you configure the jvmOptions property in the following resources:

  • Kafka.spec.kafka

  • Kafka.spec.zookeeper

  • KafkaConnect.spec

  • KafkaConnectS2I.spec

  • KafkaMirrorMaker.spec

  • KafkaMirrorMaker2.spec

  • KafkaBridge.spec

You can specify the following options in your configuration:

-Xms

Minimum initial allocation heap size when the JVM starts.

-Xmx

Maximum heap size.

-XX

Advanced runtime options for the JVM.

javaSystemProperties

Additional system properties.

gcLoggingEnabled

Enables garbage collector logging.

The full schema of jvmOptions is described in JvmOptions schema reference.

Note
The units accepted by JVM settings, such as -Xmx and -Xms, are the same units accepted by the JDK java binary in the corresponding image. Therefore, 1g or 1G means 1,073,741,824 bytes, and Gi is not a valid unit suffix. This is different from the units used for memory requests and limits, which follow the Kubernetes convention where 1G means 1,000,000,000 bytes, and 1Gi means 1,073,741,824 bytes
-Xms and -Xmx options

The default values used for -Xms and -Xmx depend on whether there is a memory request limit configured for the container.

  • If there is a memory limit, the JVM’s minimum and maximum memory is set to a value corresponding to the limit.

  • If there is no memory limit, the JVM’s minimum memory is set to 128M. The JVM’s maximum memory is not defined to allow the memory to increase as needed. This is ideal for single node environments in test and development.

Before setting -Xmx explicitly consider the following:

  • The JVM’s overall memory usage will be approximately 4 × the maximum heap, as configured by -Xmx.

  • If -Xmx is set without also setting an appropriate Kubernetes memory limit, it is possible that the container will be killed should the Kubernetes node experience memory pressure from other Pods running on it.

  • If -Xmx is set without also setting an appropriate Kubernetes memory request, it is possible that the container will be scheduled to a node with insufficient memory. In this case, the container will not start but crash immediately if -Xms is set to -Xmx, or at a later time if not.

It is recommended to:

  • Set the memory request and the memory limit to the same value

  • Use a memory request that is at least 4.5 × the -Xmx

  • Consider setting -Xms to the same value as -Xmx

In this example, the JVM uses 2 GiB (=2,147,483,648 bytes) for its heap. Its total memory usage is approximately 8GiB.

Example -Xmx and -Xms configuration
# ...
jvmOptions:
  "-Xmx": "2g"
  "-Xms": "2g"
# ...

Setting the same value for initial (-Xms) and maximum (-Xmx) heap sizes avoids the JVM having to allocate memory after startup, at the cost of possibly allocating more heap than is really needed.

Important
Containers performing lots of disk I/O, such as Kafka broker containers, require available memory for use as an operating system page cache. On such containers, the requested memory should be significantly higher than the memory used by the JVM.
-XX option

-XX options are used to configure the KAFKA_JVM_PERFORMANCE_OPTS option of Apache Kafka.

Example -XX configuration
jvmOptions:
  "-XX":
    "UseG1GC": true
    "MaxGCPauseMillis": 20
    "InitiatingHeapOccupancyPercent": 35
    "ExplicitGCInvokesConcurrent": true
JVM options resulting from the -XX configuration
-XX:+UseG1GC -XX:MaxGCPauseMillis=20 -XX:InitiatingHeapOccupancyPercent=35 -XX:+ExplicitGCInvokesConcurrent -XX:-UseParNewGC
Note
When no -XX options are specified, the default Apache Kafka configuration of KAFKA_JVM_PERFORMANCE_OPTS is used.
javaSystemProperties

javaSystemProperties are used to configure additional Java system properties, such as debugging utilities.

Example javaSystemProperties configuration
jvmOptions:
  javaSystemProperties:
    - name: javax.net.debug
      value: ssl

11.1.10. Garbage collector logging

The jvmOptions property also allows you to enable and disable garbage collector (GC) logging. GC logging is disabled by default. To enable it, set the gcLoggingEnabled property as follows:

Example GC logging configuration
# ...
jvmOptions:
  gcLoggingEnabled: true
# ...

11.2. Schema properties

11.2.1. Kafka schema reference

Property Description

spec

The specification of the Kafka and ZooKeeper clusters, and Topic Operator.

KafkaSpec

status

The status of the Kafka and ZooKeeper clusters, and Topic Operator.

KafkaStatus

11.2.2. KafkaSpec schema reference

Used in: Kafka

Property Description

kafka

Configuration of the Kafka cluster.

KafkaClusterSpec

zookeeper

Configuration of the ZooKeeper cluster.

ZookeeperClusterSpec

entityOperator

Configuration of the Entity Operator.

EntityOperatorSpec

clusterCa

Configuration of the cluster certificate authority.

CertificateAuthority

clientsCa

Configuration of the clients certificate authority.

CertificateAuthority

cruiseControl

Configuration for Cruise Control deployment. Deploys a Cruise Control instance when specified.

CruiseControlSpec

jmxTrans

Configuration for JmxTrans. When the property is present a JmxTrans deployment is created for gathering JMX metrics from each Kafka broker. For more information see JmxTrans GitHub.

JmxTransSpec

kafkaExporter

Configuration of the Kafka Exporter. Kafka Exporter can provide additional metrics, for example lag of consumer group at topic/partition.

KafkaExporterSpec

maintenanceTimeWindows

A list of time windows for maintenance tasks (that is, certificates renewal). Each time window is defined by a cron expression.

string array

11.2.3. KafkaClusterSpec schema reference

Used in: KafkaSpec

Configures a Kafka cluster.

listeners

Use the listeners property to configure listeners to provide access to Kafka brokers.

Example configuration of a plain (unencrypted) listener without authentication
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
spec:
  kafka:
    # ...
    listeners:
      - name: plain
        port: 9092
        type: internal
        tls: false
    # ...
  zookeeper:
    # ...
config

Use the config properties to configure Kafka broker options as keys.

Standard Apache Kafka configuration may be provided, restricted to those properties not managed directly by Strimzi.

Configuration options that cannot be configured relate to:

  • Security (Encryption, Authentication, and Authorization)

  • Listener configuration

  • Broker ID configuration

  • Configuration of log data directories

  • Inter-broker communication

  • ZooKeeper connectivity

The values can be one of the following JSON types:

  • String

  • Number

  • Boolean

You can specify and configure the options listed in the Apache Kafka documentation with the exception of those options that are managed directly by Strimzi. Specifically, all configuration options with keys equal to or starting with one of the following strings are forbidden:

  • listeners

  • advertised.

  • broker.

  • listener.

  • host.name

  • port

  • inter.broker.listener.name

  • sasl.

  • ssl.

  • security.

  • password.

  • principal.builder.class

  • log.dir

  • zookeeper.connect

  • zookeeper.set.acl

  • authorizer.

  • super.user

When a forbidden option is present in the config property, it is ignored and a warning message is printed to the Cluster Operator log file. All other supported options are passed to Kafka.

There are exceptions to the forbidden options. For client connection using a specific cipher suite for a TLS version, you can configure allowed ssl properties. You can also configure the zookeeper.connection.timeout.ms property to set the maximum time allowed for establishing a ZooKeeper connection.

Example Kafka broker configuration
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
    config:
      num.partitions: 1
      num.recovery.threads.per.data.dir: 1
      default.replication.factor: 3
      offsets.topic.replication.factor: 3
      transaction.state.log.replication.factor: 3
      transaction.state.log.min.isr: 1
      log.retention.hours: 168
      log.segment.bytes: 1073741824
      log.retention.check.interval.ms: 300000
      num.network.threads: 3
      num.io.threads: 8
      socket.send.buffer.bytes: 102400
      socket.receive.buffer.bytes: 102400
      socket.request.max.bytes: 104857600
      group.initial.rebalance.delay.ms: 0
      ssl.cipher.suites: "TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384"
      ssl.enabled.protocols: "TLSv1.2"
      ssl.protocol: "TLSv1.2"
      zookeeper.connection.timeout.ms: 6000
    # ...
brokerRackInitImage

When rack awareness is enabled, Kafka broker pods use init container to collect the labels from the Kubernetes cluster nodes. The container image used for this container can be configured using the brokerRackInitImage property. When the brokerRackInitImage field is missing, the following images are used in order of priority:

  1. Container image specified in STRIMZI_DEFAULT_KAFKA_INIT_IMAGE environment variable in the Cluster Operator configuration.

  2. quay.io/strimzi/operator:0.24.0 container image.

Example brokerRackInitImage configuration
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    # ...
    rack:
      topologyKey: topology.kubernetes.io/zone
    brokerRackInitImage: my-org/my-image:latest
    # ...
Note
Overriding container images is recommended only in special situations, where you need to use a different container registry. For example, because your network does not allow access to the container registry used by Strimzi. In this case, you should either copy the Strimzi images or build them from the source. If the configured image is not compatible with Strimzi images, it might not work properly.
logging

Kafka has its own configurable loggers:

  • log4j.logger.org.I0Itec.zkclient.ZkClient

  • log4j.logger.org.apache.zookeeper

  • log4j.logger.kafka

  • log4j.logger.org.apache.kafka

  • log4j.logger.kafka.request.logger

  • log4j.logger.kafka.network.Processor

  • log4j.logger.kafka.server.KafkaApis

  • log4j.logger.kafka.network.RequestChannel$

  • log4j.logger.kafka.controller

  • log4j.logger.kafka.log.LogCleaner

  • log4j.logger.state.change.logger

  • log4j.logger.kafka.authorizer.logger

Kafka uses the Apache log4j logger implementation.

Use the logging property to configure loggers and logger levels.

You can set the log levels by specifying the logger and level directly (inline) or use a custom (external) ConfigMap. If a ConfigMap is used, you set logging.valueFrom.configMapKeyRef.name property to the name of the ConfigMap containing the external logging configuration. Inside the ConfigMap, the logging configuration is described using log4j.properties. Both logging.valueFrom.configMapKeyRef.name and logging.valueFrom.configMapKeyRef.key properties are mandatory. A ConfigMap using the exact logging configuration specified is created with the custom resource when the Cluster Operator is running, then recreated after each reconciliation. If you do not specify a custom ConfigMap, default logging settings are used. If a specific logger value is not set, upper-level logger settings are inherited for that logger. For more information about log levels, see Apache logging services.

Here we see examples of inline and external logging.

Inline logging
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
spec:
  # ...
  kafka:
    # ...
    logging:
      type: inline
      loggers:
        kafka.root.logger.level: "INFO"
  # ...
External logging
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
spec:
  # ...
  logging:
    type: external
    valueFrom:
      configMapKeyRef:
        name: customConfigMap
        key: kafka-log4j.properties
  # ...

Any available loggers that are not configured have their level set to OFF.

If Kafka was deployed using the Cluster Operator, changes to Kafka logging levels are applied dynamically.

If you use external logging, a rolling update is triggered when logging appenders are changed.

Garbage collector (GC)

Garbage collector logging can also be enabled (or disabled) using the jvmOptions property.

KafkaClusterSpec schema properties
Property Description

version

The kafka broker version. Defaults to 2.8.0. Consult the user documentation to understand the process required to upgrade or downgrade the version.

string

replicas

The number of pods in the cluster.

integer

image

The docker image for the pods. The default value depends on the configured Kafka.spec.kafka.version.

string

listeners

Configures listeners of Kafka brokers.

GenericKafkaListener array

config

Kafka broker config properties with the following prefixes cannot be set: listeners, advertised., broker., listener., host.name, port, inter.broker.listener.name, sasl., ssl., security., password., principal.builder.class, log.dir, zookeeper.connect, zookeeper.set.acl, zookeeper.ssl, zookeeper.clientCnxnSocket, authorizer., super.user, cruise.control.metrics.topic, cruise.control.metrics.reporter.bootstrap.servers (with the exception of: zookeeper.connection.timeout.ms, ssl.cipher.suites, ssl.protocol, ssl.enabled.protocols,cruise.control.metrics.topic.num.partitions, cruise.control.metrics.topic.replication.factor, cruise.control.metrics.topic.retention.ms,cruise.control.metrics.topic.auto.create.retries, cruise.control.metrics.topic.auto.create.timeout.ms,cruise.control.metrics.topic.min.insync.replicas).

map

storage

Storage configuration (disk). Cannot be updated. The type depends on the value of the storage.type property within the given object, which must be one of [ephemeral, persistent-claim, jbod].

EphemeralStorage, PersistentClaimStorage, JbodStorage

authorization

Authorization configuration for Kafka brokers. The type depends on the value of the authorization.type property within the given object, which must be one of [simple, opa, keycloak, custom].

KafkaAuthorizationSimple, KafkaAuthorizationOpa, KafkaAuthorizationKeycloak, KafkaAuthorizationCustom

rack

Configuration of the broker.rack broker config.

Rack

brokerRackInitImage

The image of the init container used for initializing the broker.rack.

string

livenessProbe

Pod liveness checking.

Probe

readinessProbe

Pod readiness checking.

Probe

jvmOptions

JVM Options for pods.

JvmOptions

jmxOptions

JMX Options for Kafka brokers.

KafkaJmxOptions

resources

CPU and memory resources to reserve. For more information, see the external documentation for core/v1 resourcerequirements.

ResourceRequirements

metricsConfig

Metrics configuration. The type depends on the value of the metricsConfig.type property within the given object, which must be one of [jmxPrometheusExporter].

JmxPrometheusExporterMetrics

logging

Logging configuration for Kafka. The type depends on the value of the logging.type property within the given object, which must be one of [inline, external].

InlineLogging, ExternalLogging

template

Template for Kafka cluster resources. The template allows users to specify how are the StatefulSet, Pods and Services generated.

KafkaClusterTemplate

11.2.4. GenericKafkaListener schema reference

Used in: KafkaClusterSpec

Configures listeners to connect to Kafka brokers within and outside Kubernetes.

You configure the listeners in the Kafka resource.

Example Kafka resource showing listener configuration
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-cluster
spec:
  kafka:
    #...
    listeners:
      - name: plain
        port: 9092
        type: internal
        tls: false
      - name: tls
        port: 9093
        type: internal
        tls: true
        authentication:
          type: tls
      - name: external1
        port: 9094
        type: route
        tls: true
      - name: external2
        port: 9095
        type: ingress
        tls: true
        authentication:
          type: tls
        configuration:
          bootstrap:
            host: bootstrap.myingress.com
          brokers:
          - broker: 0
            host: broker-0.myingress.com
          - broker: 1
            host: broker-1.myingress.com
          - broker: 2
            host: broker-2.myingress.com
    #...
listeners

You configure Kafka broker listeners using the listeners property in the Kafka resource. Listeners are defined as an array.

Example listener configuration
listeners:
  - name: plain
    port: 9092
    type: internal
    tls: false

The name and port must be unique within the Kafka cluster. The name can be up to 25 characters long, comprising lower-case letters and numbers. Allowed port numbers are 9092 and higher with the exception of ports 9404 and 9999, which are already used for Prometheus and JMX.

By specifying a unique name and port for each listener, you can configure multiple listeners.

type

The type is set as internal, or for external listeners, as route, loadbalancer, nodeport or ingress.

internal

You can configure internal listeners with or without encryption using the tls property.

Example internal listener configuration
#...
spec:
  kafka:
    #...
    listeners:
      #...
      - name: plain
        port: 9092
        type: internal
        tls: false
      - name: tls
        port: 9093
        type: internal
        tls: true
        authentication:
          type: tls
    #...
route

Configures an external listener to expose Kafka using OpenShift Routes and the HAProxy router.

A dedicated Route is created for every Kafka broker pod. An additional Route is created to serve as a Kafka bootstrap add