Reactor Kafka Application Stops Unexpectedly Without Errors During High-Volume Consumption

Question

I am currently working on a reactor Kafka application designed to process millions of Kafka records per day. Despite careful configuration and the implementation of a retry mechanism, I am encountering an issue where the consumption of records unexpectedly halts without any apparent reason. Notably, there are no out-of-memory errors, application crashes, or reactor Kafka debug logs indicating what might be wrong.

Configuration Details:

• maxDeferredCommits: Set to 8000
• maxCommitAttempts: Configured to 200
• commitRetryInterval: 25ms
• Retry Mechanism: Implemented via .retryWhen() with a duration strategy and limited to 3 retries.

Despite these settings, the application's consumption process stops randomly, and I cannot determine the cause due to the lack of error messages or logs.

Question 1: Could leveraging .repeat() in some way help mitigate this issue of sudden cessation in consumption? If so, how is it different from the .retryWhen operator in this context?

Question 2: Are there additional strategies or configurations I should consider to diagnose or resolve the underlying cause of this unexpected halt in consumption?

I am looking for insights or suggestions on what might be causing this behavior and how to prevent it. Any guidance on further debugging steps or adjustments to my configuration that could help maintain stable consumption would be greatly appreciated.