Kafka or Rabbitmq message brokers can be deployed conditionally as part of Helm release using rabbitmq.enabled or kafka.enabled. The default values for Kafka and Rabbitmq have been set up to use message brokers provided with the Helm chart. It is possible to disable message broker deployments and use external Kafka or Rabbitmq by overriding global.kafka or global.rabbitmq values, i.e. global.kafka.brokers (Multiple brokers can be provided in a comma-separated list, e.g. host1:port1,host2:port2),global.kafka.zkNodes (External Zookeeper nodes).
In the milestone 7.1.0-M13 you will find the following :
: messages used to transfer data from the runtime bundle to query and audit service can be partitioned. There will be as many instances of the consumer as the number of partitions. Each instance will consume data from a single partition. In Helm deployment, partitioning can be enabled by providing (partitionCount defines how many partitions will be used and the Helm deployment will create that many replicas of query service using a StatefulSet rather than a Deployment):
: in addition to Rabbit MQ binder (that remains the default one), the example docker images generated by are also embedding Kafka binder. In order to use it instead of Rabbit MQ binder, pass the following during Helm deployment:
