This is a highly configurable Dockerized Confluent Platform Kafka Image.
The container can be configured via environment variables where any Kafka property can be set after kafka_ with underscore instead of period ('.') used. For example:
| Environment Variable | Property | Default |
|---|---|---|
| kafka_auto_create_topics_enable | auto.create.topics.enable | true |
| kafka_broker_id | broker.id | 1 |
| kafka_delete_topic_enable | delete.topic.enable | true |
| kafka_log_cleaner_enable | log.cleaner.enable | true |
| kafka_log_retention_hours | log.retention.hours | 168 |
| kafka_num_partitions | num.partitions | 1 |
| kafka_num_recovery_threads_per_data_dir | num.recovery.threads.per.data.dir | 1 |
| kafka_port | port | 9092 |
| kafka_zookeeper_connect | zookeeper.connect | ZOOKEEPER_PORT_2181_TCP_ADDR:ZOOKEEPER_PORT_2181_TCP_PORT |
A few sensible, and a few opinionated values, have been set as given above. A basic invocation using a link to a local ZooKeeper instance would be:
docker run --rm --name kafka -p 9092:9092 --link zookeeper:zookeeper cgswong/confluent-kafka
The config directory, /etc/kafka, is exposed for mounting to your local host. This facilitates using your own configuration file, kafka.properties instead of injecting into the container, while still take advantage of variable substitution. Also, you can download a remote properties file by setting the environment variable kafka_cfg_url to the location of the file.
This image is a fork of the main ConfluentInc Platform. As I do contribute back you may see some of these features included, however, some may not be included or take longer to be merged. I created this version not only to contribute back to a really good project, but also to learn, and have a version which was more robust (production ready since this is being used as such), with a more timely update cycle (for the time being anyways).
docker pull cgswong/confluent-kafka