A library that provides in-memory instances of both Kafka and Confluent Schema Registry to run your tests against.
Relies on the embedded-kafka library.
embedded-kafka-schema-registry is available on Maven Central, compiled for both Scala 2.11 and 2.12.
Currently there's no support for Scala 2.13-Mx as Confluent artifacts are not published for these versions.
Versions match the version of Confluent Schema Registry they're built against.
- In your
build.sbt
file add the following resolver:resolvers += "confluent" at "https://packages.confluent.io/maven/"
- In your
build.sbt
file add the following dependency:"io.github.embeddedkafka" %% "embedded-kafka-schema-registry" % "5.2.2" % "test"
- Have your class extend the
EmbeddedKafka
trait (from thenet.manub.embeddedkafka.schemaregistry
package). - Enclose the code that needs a running instance of Kafka within the
withRunningKafka
closure. - Provide an implicit
EmbeddedKafkaConfigImpl
(from the same package mentioned before).
class MySpec extends WordSpec with EmbeddedKafka {
"runs with embedded kafka and Schema Registry" should {
"work" in {
implicit val config = EmbeddedKafkaConfigImpl()
withRunningKafka {
// ... code goes here
}
}
}
}
- In-memory Zookeeper, Kafka, and Schema Registry will be instantiated respectively on port 6000, 6001, and 6002 and automatically shutdown at the end of the test.
The net.manub.embeddedkafka.avro.schemaregistry
package object provides useful implicit converters for testing with Avro and Schema Registry.
- For most of the cases have your class extend the
EmbeddedKafkaStreamsAllInOne
trait (from thenet.manub.embeddedkafka.schemaregistry.streams
package). This offers both streams management and easy creation of consumers for asserting resulting messages in output/sink topics. - If you only want to use the streams management without the test consumers just have the class extend the
EmbeddedKafkaStreams
trait (from the same package mentioned before). - Build your own
Topology
and userunStreams
to test it. - Have a look at the example test.