Mastering Kafka Streams and ksqlDB

Appendix B. ksqlDB Configuration

ksqlDB accepts most Kafka Streams and Kafka Client (i.e., producer and consumer) configurations. The recommended pattern is to prefix all Kafka Streams and Kafka Client configs with ksql.streams. For example, if you want to configure a record cache using the cache.max.bytes.buffering parameter in Kafka Streams, you would set the ksql.streams.cache.max.bytes.buffering parameter in your server.properties file. Furthermore, if you want to configure the auto.offset.reset Kafka Consumer configuration, this would also be prefixed the same way, and would become ksql.streams.auto.offset.reset. Technically, the prefix is optional, but comes as a recommendation from the creators of ksqlDB (Confluent).

In addition to the standard Kafka Streams and Kafka Client configs, ksqlDB also allows you to specify a Kafka Connect configuration if you are using ksqlDB’s data integration features (e.g., whenever you are executing a CREATE {SOURCE|SINK} CONNECTOR statement). We’ve already discussed this in “Configuring Connect Workers”, so please see that section for more detail.

Finally, there are several configurations that are specific to ksqlDB. We have grouped some of the most important configurations into two categories: query configurations and server configurations. This page should be used as a starting point for configuring ksqlDB. Please refer to the official documentation for a full list of ksqlDB configuration parameters.¹

Query Configurations

The following ...