Main Changes since 0.20.x
Note: This version of Strimzi supports Kubernetes 1.16 and higher.
- Add support for Kafka 2.7.0
- Add support for declarative management of connector plugins in Kafka Connect CR
- Add
inter.broker.protocol.versionto the default configuration in example YAMLs - Add support for
secretPrefixproperty for User Operator to prefix all secret names created from KafkaUser resource. - Allow configuring labels and annotations for Cluster CA certificate secrets
- Add the JAAS configuration string in the sasl.jaas.config property to the generated secrets for KafkaUser with SCRAM-SHA-512 authentication.
- Strimzi
test-containerhas been renamed tostrimzi-test-containerto make the name more clear - Updated the CPU usage metric in the Kafka, ZooKeeper and Cruise Control dashboards to include the CPU kernel time (other than the current user time)
- Allow disabling ownerReference on CA secrets
- Make it possible to run Strimzi operators and operands with read-only root filesystem
- Move from Docker Hub to Quay.io as our container registry
- Add possibility to configure DeploymentStrategy for Kafka Connect, Kafka Mirror Maker (1 and 2), and Kafka Bridge
- Support passing metrics configuration as an external ConfigMap
- Enable CORS configuration for Cruise Control
- Add support for rolling individual Kafka or ZooKeeper pods through the Cluster Operator using an annotation
- Add support for Topology Spread Constraints in Pod templates
- Make Kafka
cluster-id(KIP-78) available on Kafka CRD status
All changes can be found under the 0.21.0 milestone.
Deprecations and removals
- The
metricsfield in the Strimzi custom resources has been deprecated and will be removed in the future. For configuring metrics, use the newmetricsConfigfield and pass the configuration via ConfigMap.
!!!!! IMPORTANT - Update 17th Jan 2021 !!!!!
There is a bug in the new feature for configuring metrics from ConfigMap instead of directly in the custom resource. When used, it in some cases does not properly configure the network policies for the port 9404 used for the metrics. As a workaround, either keep using the old configuration or create manually an additional network policy with your custom name for the port 9404. This bug will be fixed in 0.21.1.