Version 1.0.0 of Kafka for Junit has been released. It features a couple of small improvements and bugfixes but most notably works now with Kafka 1.1.x and adds two significant epics Multiple Brokers and Failure Modes as well as an extended queriability when reading from or observing a topic, since both operations now accept a filter on record headers. With this release, it is possible to provision an embedded cluster with multiple brokers. This led to the possiblity to introduce failure modes on the broker-level as well:
EmbeddedKafkaCluster provides a set of methods that allow you to disconnect specific brokers or disconnect as many brokers as required to fall below the minimum size of the In-Sync-Replica Set of a dedicated topic. This enables you to test if your Kafka-enabled component or application behaves correctly in the presence of broker outages.