A setup for Kafka message broker based on confluent
To start the cluster run following command:
docker-compose upAlternatively docker stack can be used.
The following environment variables are used:
- KAFKA_BROKER_ID The unique id for each kafka instance
- KAFKA_ZOOKEEPER_CONNECT: List of all Zookeeper instances with client port
- KAFKA_ADVERTISED_LISTENERS: Listener to publish for clients (default: INTERNAL://{service_name}:9092)
- KAFKA_INTER_BROKER_LISTENER_NAME : Name of listener to be used (default: INTERNAL)
- KAFKA_LISTENER_SECURITY_PROTOCOL_MAP : Map between listener an security protocols
- KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR : Replication factor for offset topic (default: 1) Longer explanations can be found at the Kafka documentation
The following environment variables are used:
- ZOOKEEPER_SERVER_ID The unique id for each zookeeper instance
- ZOOKEEPER_CLIENT_PORT Set the client port (default: 2181)
- ZOOKEEPER_TICK_TIME Set the length of a tick, which is base for heartbeats and timeouts (default: 2000 ms)
- ZOOKEEPER_INIT_LIMIT Amount of time to allow followers to connect (default: 5 ticks)
- ZOOKEEPER_SYNC_LIMIT Amount of time to allow followers to sync (default: 2 ticks)
- ZOOKEEPER_SERVERS List of all Zookeeper instances with ports Longer explanations can be found at the zookeeper documentation
The project is licensed under the Apache 2 license.