Configuring Strimzi (0.40.0)

Use the replicas property to configure replicas.

The type of replication depends on the resource.

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.

You can incorporate SSL configuration and cipher suite specifications to further secure TLS-based communication between your client application and a Kafka cluster. In addition to the standard TLS configuration, you can specify a supported TLS version and enable cipher suites in the configuration for the Kafka broker. You can also add the configuration to your clients if you wish to limit the TLS versions and cipher suites they use. The configuration on the client must only use protocols and cipher suites that are enabled on the broker.

A cipher suite is a set of security mechanisms for secure connection and data transfer. For example, the cipher suite TLS_AES_256_GCM_SHA384 is composed of the following mechanisms, which are used in conjunction with the TLS protocol:

The combination is encapsulated in the TLS_AES_256_GCM_SHA384 cipher suite specification.

The ssl.enabled.protocols property specifies the available TLS versions that can be used for secure communication between the cluster and its clients. The ssl.protocol property sets the default TLS version for all connections, and it must be chosen from the enabled protocols. Use the ssl.endpoint.identification.algorithm property to enable or disable hostname verification (configurable only in components based on Kafka clients - Kafka Connect, MirrorMaker 1/2, and Kafka Bridge).

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:

If certificates are stored in the same secret, it can be listed multiple times.

If you want to enable TLS encryption, but use the default set of public certification authorities shipped with Java, you can specify trustedCertificates as an empty array:

For information on configuring mTLS authentication, see the KafkaClientAuthenticationTls schema reference.

Configure resource requests and limits to control resources for Strimzi containers. You can specify requests and limits for memory and cpu resources. The requests should be enough to ensure a stable performance of Kafka.

How you configure resources in a production environment depends on a number of factors. For example, applications are likely to be sharing resources in your Kubernetes cluster.

For Kafka, the following aspects of a deployment can impact the resources you need:

The values specified for resource requests are reserved and always available to the container. Resource limits specify the maximum resources that can be consumed by a given container. The amount between the request and limit is not reserved and might not be always available. A container can use the resources up to the limit only when they are available. Resource limits are temporary and can be reallocated.

If you set limits without requests or vice versa, Kubernetes uses the same value for both. Setting equal requests and limits for resources guarantees quality of service, as Kubernetes will not kill containers unless they exceed their limits.

You can configure resource requests and limits for one or more supported resources.

Resource requests and limits for the Topic Operator and User Operator are set in the Kafka resource.

If the resource request is for more than the available free resources in the Kubernetes cluster, the pod is not scheduled.

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:

Configuring the image property for Kafka, Kafka Connect, and Kafka MirrorMaker

Kafka, Kafka Connect, 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:

These environment variables contain mappings between Kafka versions and corresponding images. The mappings are used together with the image and version properties to determine the image used:

The image and version for the components can be configured in the following properties:

Configuring the image property in other resources

For the image property in the custom resources for other components, the given value is used during deployment. If the image property is not set, the container image specified as an environment variable in the Cluster Operator configuration is used. If an image name is not defined in the Cluster Operator configuration, then a default value is used.

For more information on image environment variables, see Configuring the Cluster Operator.

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:

For more information about the livenessProbe and readinessProbe options, see the Probe schema reference.

Use the metricsConfig property to enable and configure Prometheus metrics.

The metricsConfig property contains a reference to a ConfigMap that has additional configurations 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.

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.

The following Strimzi components run inside a Java Virtual Machine (JVM):

To optimize their performance on different platforms and architectures, you configure the jvmOptions property in the following resources:

You can specify the following options in your configuration:

In addition to setting memory request and limit values for your containers, you can use the -Xms and -Xmx JVM options to set specific heap sizes for your JVM. Use the -Xms option to set an initial heap size and the -Xmx option to set a maximum heap size.

Specify heap size to have more control over the memory allocated to your JVM. Heap sizes should make the best use of a container’s memory limit (and request) without exceeding it. Heap size and any other memory requirements need to fit within a specified memory limit. If you don’t specify heap size in your configuration, but you configure a memory resource limit (and request), the Cluster Operator imposes default heap sizes automatically. The Cluster Operator sets default maximum and minimum heap values based on a percentage of the memory resource configuration.

The following table shows the default heap values.

If a memory limit (and request) is not specified, a JVM’s minimum heap size is set to 128M. The JVM’s maximum heap size is not defined to allow the memory to increase as needed. This is ideal for single node environments in test and development.

Setting an appropriate memory request can prevent the following:

In this example, the JVM uses 2 GiB (=2,147,483,648 bytes) for its heap. Total JVM memory usage can be a lot more than the maximum heap size.

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.

-XX options are used to configure the KAFKA_JVM_PERFORMANCE_OPTS option of Apache Kafka.

javaSystemProperties are used to configure additional Java system properties, such as debugging utilities.

For more information about the jvmOptions, see the JvmOptions schema reference.

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:

Use the listeners property to configure listeners to provide access to Kafka brokers.

Use the config properties to configure Kafka broker options as keys.

The values can be one of the following JSON types:

Exceptions

You can specify and configure the options listed in the Apache Kafka documentation.

However, Strimzi takes care of configuring and managing options related to the following, which cannot be changed:

Properties with the following prefixes cannot be set:

If the config property contains an option that cannot be changed, it is disregarded, and a warning message is logged to the Cluster Operator log file. All other supported options are forwarded to Kafka, including the following exceptions to the options configured by Strimzi:

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:

Kafka has its own configurable loggers, which include the following:

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. The inline logging specifies the root logger level. You can also set log levels for specific classes or loggers by adding them to the loggers property.

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 logging can also be enabled (or disabled) using the jvmOptions property.

You configure Kafka broker listeners using the listeners property in the Kafka resource. Listeners are defined as an array.

The name and port must be unique within the Kafka cluster. By specifying a unique name and port for each listener, you can configure multiple listeners. The name can be up to 25 characters long, comprising lower-case letters and numbers.

The port number is the port used in the Kafka cluster, which might not be the same port used for access by a client.

For client connection, use the address and port for the bootstrap service of the listener. You can retrieve this from the status of the Kafka resource.

The type is set as internal, or for external listeners, as route, loadbalancer, nodeport, ingress or cluster-ip. You can also configure a cluster-ip listener, a type of internal listener you can use to build custom access mechanisms.

The TLS property is required.

By default, TLS encryption is not enabled. To enable it, set the tls property to true.

For route and ingress type listeners, TLS encryption must be enabled.

Authentication for the listener can be specified as:

Use networkPolicyPeers to configure network policies that restrict access to a listener at the network level. The following example shows a networkPolicyPeers configuration for a plain and a tls listener.

In the following example:

The syntax of the networkPolicyPeers property is the same as the from property in NetworkPolicy resources.

Listener configuration specified using listenerConfig is prefixed with listener.name.<listener_name>-<port>. For example, sasl.enabled.mechanisms becomes listener.name.<listener_name>-<port>.sasl.enabled.mechanisms.

Secrets are mounted to /opt/kafka/custom-authn-secrets/custom-listener-<listener_name>-<port>/<secret_name> in the Kafka broker nodes' containers.

For example, the mounted secret (example) in the example configuration would be located at /opt/kafka/custom-authn-secrets/custom-listener-oauth-bespoke-9093/example.

You can set a custom principal builder in the Kafka cluster configuration. However, the principal builder is subject to the following requirements:

Custom principal builders must support peer certificates for authentication, as Strimzi uses these to manage the Kafka cluster.

A custom OAuth principal builder might be identical or very similar to the Strimzi OAuth principal builder.

The following example shows a custom principal builder that satisfies the OAuth requirements of Strimzi.

The type property is a discriminator that distinguishes use of the KafkaListenerAuthenticationCustom type from KafkaListenerAuthenticationTls, KafkaListenerAuthenticationScramSha512, KafkaListenerAuthenticationOAuth. It must have the value custom for the type KafkaListenerAuthenticationCustom.

The brokerCertChainAndKey property is only used with listeners that have TLS encryption enabled. You can use the property to provide your own Kafka listener certificates.

When the certificate or key in the brokerCertChainAndKey secret is updated, the operator will automatically detect it in the next reconciliation and trigger a rolling update of the Kafka brokers to reload the certificate.

The externalTrafficPolicy property is used with loadbalancer and nodeport listeners. When exposing Kafka outside of Kubernetes you can choose Local or Cluster. Local avoids hops to other nodes and preserves the client IP, whereas Cluster does neither. The default is Cluster.

The loadBalancerSourceRanges property is only used with loadbalancer listeners. When exposing Kafka outside of Kubernetes use source ranges, in addition to labels and annotations, to customize how a service is created.

The class property is only used with ingress listeners. You can configure the Ingress class using the class property.

The preferredNodePortAddressType property is only used with nodeport listeners.

Use the preferredNodePortAddressType property in your listener configuration to specify the first address type checked as the node address. This property is useful, for example, if your deployment does not have DNS support, or you only want to expose a broker internally through an internal DNS or IP address. If an address of this type is found, it is used. If the preferred address type is not found, Strimzi proceeds through the types in the standard order of priority:

The useServiceDnsDomain property is only used with internal and cluster-ip listeners. It defines whether the fully-qualified DNS names that include the cluster service suffix (usually .cluster.local) are used. With useServiceDnsDomain set as false, the advertised addresses are generated without the service suffix; for example, my-cluster-kafka-0.my-cluster-kafka-brokers.myproject.svc. With useServiceDnsDomain set as true, the advertised addresses are generated with the service suffix; for example, my-cluster-kafka-0.my-cluster-kafka-brokers.myproject.svc.cluster.local. Default is false.

If your Kubernetes cluster uses a different service suffix than .cluster.local, you can configure the suffix using the KUBERNETES_SERVICE_DNS_DOMAIN environment variable in the Cluster Operator configuration.

You can specify alternative names for the bootstrap service. The names are added to the broker certificates and can be used for TLS hostname verification. The alternativeNames property is applicable to all types of listeners.

The host property is used with route and ingress listeners to specify the hostnames used by the bootstrap and per-broker services.

A host property value is mandatory for ingress listener configuration, as the Ingress controller does not assign any hostnames automatically. Make sure that the hostnames resolve to the Ingress endpoints. Strimzi will not perform any validation that the requested hosts are available and properly routed to the Ingress endpoints.

By default, route listener hosts are automatically assigned by OpenShift. However, you can override the assigned route hosts by specifying hosts.

Strimzi does not perform any validation that the requested hosts are available. You must ensure that they are free and can be used.

By default, the port numbers used for the bootstrap and broker services are automatically assigned by Kubernetes. You can override the assigned node ports for nodeport listeners by specifying the requested port numbers.

Strimzi does not perform any validation on the requested ports. You must ensure that they are free and available for use.

Use the loadBalancerIP property to request a specific IP address when creating a loadbalancer. Use this property when you need to use a loadbalancer with a specific IP address. The loadBalancerIP field is ignored if the cloud provider does not support the feature.

Use the annotations property to add annotations to Kubernetes resources related to the listeners. You can use these annotations, for example, to instrument DNS tooling such as External DNS, which automatically assigns DNS names to the loadbalancer services.

A list of user principals treated as super users, so that they are always allowed without querying ACL rules.

The type property is a discriminator that distinguishes use of the KafkaAuthorizationSimple type from KafkaAuthorizationOpa, KafkaAuthorizationKeycloak, KafkaAuthorizationCustom. It must have the value simple for the type KafkaAuthorizationSimple.

The URL used to connect to the Open Policy Agent server. The URL has to include the policy which will be queried by the authorizer. Required.

Defines whether a Kafka client should be allowed or denied by default when the authorizer fails to query the Open Policy Agent, for example, when it is temporarily unavailable. Defaults to false - all actions will be denied.

Initial capacity of the local cache used by the authorizer to avoid querying the Open Policy Agent for every request. Defaults to 5000.

Maximum capacity of the local cache used by the authorizer to avoid querying the Open Policy Agent for every request. Defaults to 50000.

The expiration of the records kept in the local cache to avoid querying the Open Policy Agent for every request. Defines how often the cached authorization decisions are reloaded from the Open Policy Agent server. In milliseconds. Defaults to 3600000 milliseconds (1 hour).

Trusted certificates for TLS connection to the OPA server.

A list of user principals treated as super users, so that they are always allowed without querying the open Policy Agent policy.

The type property is a discriminator that distinguishes use of the KafkaAuthorizationOpa type from KafkaAuthorizationSimple, KafkaAuthorizationKeycloak, KafkaAuthorizationCustom. It must have the value opa for the type KafkaAuthorizationOpa.

(Required) Java class that implements the org.apache.kafka.server.authorizer.Authorizer interface to support custom ACLs.

A list of user principals treated as super users, so that they are always allowed without querying ACL rules.

You can add configuration for initializing the custom authorizer using Kafka.spec.kafka.config.

In addition to the Kafka custom resource configuration, the JAR file containing the custom authorizer class along with its dependencies must be available on the classpath of the Kafka broker.

The Strimzi Maven build process provides a mechanism to add custom third-party libraries to the generated Kafka broker container image by adding them as dependencies in the pom.xml file under the docker-images/kafka/kafka-thirdparty-libs directory. The directory contains different folders for different Kafka versions. Choose the appropriate folder. Before modifying the pom.xml file, the third-party library must be available in a Maven repository, and that Maven repository must be accessible to the Strimzi build process.

Custom authorization can make use of group membership information extracted from the JWT token during authentication when using oauth authentication and configuring groupsClaim configuration attribute. Groups are available on the OAuthKafkaPrincipal object during authorize() call as follows:

The type property is a discriminator that distinguishes use of the KafkaAuthorizationCustom type from KafkaAuthorizationSimple, KafkaAuthorizationOpa, KafkaAuthorizationKeycloak. It must have the value custom for the type KafkaAuthorizationCustom.

When rack awareness is configured, Strimzi will set broker.rack configuration for each Kafka broker. The broker.rack configuration assigns a rack ID to each broker. When broker.rack is configured, Kafka brokers will spread partition replicas across as many different racks as possible. When replicas are spread across multiple racks, the probability that multiple replicas will fail at the same time is lower than if they would be in the same rack. Spreading replicas improves resiliency, and is important for availability and reliability. To enable rack awareness in Kafka, add the rack option to the .spec.kafka section of the Kafka custom resource as shown in the example below.

When rack awareness is enabled in the Kafka custom resource, Strimzi will automatically add the Kubernetes preferredDuringSchedulingIgnoredDuringExecution affinity rule to distribute the Kafka brokers across the different racks. However, the preferred rule does not guarantee that the brokers will be spread. Depending on your exact Kubernetes and Kafka configurations, you should add additional affinity rules or configure topologySpreadConstraints for both ZooKeeper and Kafka to make sure the nodes are properly distributed accross as many racks as possible. For more information see Configuring pod scheduling.

Rack awareness can also be used in consumers to fetch data from the closest replica. This is useful for reducing the load on your network when a Kafka cluster spans multiple datacenters and can also reduce costs when running Kafka in public clouds. However, it can lead to increased latency.

In order to be able to consume from the closest replica, rack awareness has to be configured in the Kafka cluster, and the RackAwareReplicaSelector has to be enabled. The replica selector plugin provides the logic that enables clients to consume from the nearest replica. The default implementation uses LeaderSelector to always select the leader replica for the client. Specify RackAwareReplicaSelector for the replica.selector.class to switch from the default implementation.

In addition to the Kafka broker configuration, you also need to specify the client.rack option in your consumers. The client.rack option should specify the rack ID in which the consumer is running. RackAwareReplicaSelector associates matching broker.rack and client.rack IDs, to find the nearest replica and consume from it. If there are multiple replicas in the same rack, RackAwareReplicaSelector always selects the most up-to-date replica. If the rack ID is not specified, or if it cannot find a replica with the same rack ID, it will fall back to the leader replica.

You can also configure Kafka Connect, MirrorMaker 2 and Kafka Bridge so that connectors consume messages from the closest replicas. You enable rack awareness in the KafkaConnect, KafkaMirrorMaker2, and KafkaBridge custom resources. The configuration does does not set affinity rules, but you can also configure affinity or topologySpreadConstraints. For more information see Configuring pod scheduling.

When deploying Kafka Connect using Strimzi, you can use the rack section in the KafkaConnect custom resource to automatically configure the client.rack option.

When deploying MirrorMaker 2 using Strimzi, you can use the rack section in the KafkaMirrorMaker2 custom resource to automatically configure the client.rack option.

When deploying Kafka Bridge using Strimzi, you can use the rack section in the KafkaBridge custom resource to automatically configure the client.rack option.

Use the hostAliases property to a specify a list of hosts and IP addresses, which are injected into the /etc/hosts file of the pod.

This configuration is especially useful for Kafka Connect or MirrorMaker when a connection outside of the cluster is also requested by users.

The type property is a discriminator that distinguishes use of the TieredStorageCustom type from other subtypes which may be added in the future. It must have the value custom for the type TieredStorageCustom.

Use the config properties to configure ZooKeeper options as keys.

The values can be one of the following JSON types:

Exceptions

You can specify and configure the options listed in the ZooKeeper documentation.

However, Strimzi takes care of configuring and managing options related to the following, which cannot be changed:

Properties with the following prefixes cannot be set:

If the config property contains an option that cannot be changed, it is disregarded, and a warning message is logged to the Cluster Operator log file. All other supported options are forwarded to ZooKeeper, including the following exceptions to the options configured by Strimzi:

ZooKeeper has a configurable logger:

ZooKeeper 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. The inline logging specifies the root logger level. You can also set log levels for specific classes or loggers by adding them to the loggers property.

Garbage collector logging can also be enabled (or disabled) using the jvmOptions property.

The Topic Operator has a configurable logger:

The Topic Operator uses the Apache log4j2 logger implementation.

Use the logging property in the entityOperator.topicOperator field of the Kafka resource Kafka resource 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 log4j2.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. The inline logging specifies the root logger level. You can also set log levels for specific classes or loggers by adding them to the loggers property.

Garbage collector logging can also be enabled (or disabled) using the jvmOptions property.

The User Operator has a configurable logger:

The User Operator uses the Apache log4j2 logger implementation.

Use the logging property in the entityOperator.userOperator field of the Kafka resource 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 log4j2.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. The inline logging specifies the rootLogger.level. You can also set log levels for specific classes or loggers by adding them to the loggers property.

Garbage collector logging can also be enabled (or disabled) using the jvmOptions property.