THIS IS A TEST INSTANCE. ALL YOUR CHANGES WILL BE LOST!!!!
...
Current state: Discuss
Discussion Thread: here
JIRA: Jira server ASF JIRA serverId 5aa69414-a9e9-3523-82ec-879b028fb15b key KAFKA-13152
...
- Whenever all the tasks belonging to a particular Stream Thread have amassed bytes more than what has been allocated to it, the underneath consumer would pause consuming from all partitions that have some data already. What this also means, is that partitions with empty buffers wouldn't be paused.
- The paused partitions would be resumed after the buffered size for all tasks in the thread are lesser than the bytes allocated.
- Whenever there is any modification in the number of Stream Threads in the topology, the buffer bytes allocation to all Stream Threads would change. This behaviour is exactly similar to
cache.max.bytes.buffering.
- As per the suggestions on the dev mailing list, we would also be renaming the config cache.max.bytes.buffering to statestore.cache.max.bytes as part of this KIP.
Public Interfaces
- Adding a new config input.buffer.max.bytes applicable at a topology level. The importance of this config would be Medium.
- Renaming cache.max.bytes.buffering to statestore.cache.max.bytes.
Compatibility, Deprecation, and Migration Plan
...