retry.backoff.ms is affected by offset.flush.interval.ms during retries #1523
Description
I'm not sure if this is an expected behavior, but I couldn't find any documented relation between these two parameters.
For example, I have a sink connector configured with retry.backoff.ms=60000, and the worker running the sink task has offset.flush.interval.ms=10000. When the task fails. the retry is triggered, but the delay is not respected - instead of waiting 60000ms, it retries after just 10000ms.
It seems like the offset flush cycle is interfering with the retry backoff mechanism.
Is this the intended behavior?
Environment:
- Kafka Connect: 7.7.2
- Kafka Connect JDBC: 10.8.4
- Kafka: 3.8.0
- Java: OpenJDK 17
- Deployment: Distributed
- OS: Alpine 3.16