The persistent storage footprint of a Kafka Streams application contains the following aspects:
- The internal topics created on the Kafka cluster side.
- The materialized state stores on the Kafka Streams application instances side.
There have been some questions about reducing these footprints, especially since many of them are not necessary. For example, there are redundant internal topics, as well as unnecessary state stores that takes up space but also affect performance. When people are pushing Streams to production with high traffic, this issue would be more common and severe. Reducing the footprint of Streams have clear benefits for reducing resource utilization of Kafka Streams applications, and also not creating pressure on broker's capacities.