Artem Bilan

Got it! So, we do agree that pause happens anyway. Only the problem that we don't `poll()` within that batch retry loop. Is that correct?

> (delay between retries) which is already greater than `max.poll.interval`. I don't think this is OK, since retry here is really blocking. So, we definitely must sleep not more than...

OK. I think you indeed are talking about a graceful shutdown. First of all I was going to reject your request, but now I see that in Spring AMQP we...

Right. We have same doubts. We will investigate how `ContainerPausingBackOffHandler` works in records mode and come back to you.

@rohan-changejar , do you mean that you are also trying to the `ContainerPausingBackOffHandler` from this discussion? Or do you mean that your back-off timeout is really longer than poll interval?...

OK! Thank you for looking into that! Apparently you are fully on board with the code. So, we have everything what could give us a graceful shutdown. Your only concern...

Has been closed accidentally by the related PR. However the fix in the PR does nothing with the `ContainerPausingBackOffHandler` support for batch itself. According to the docs https://docs.spring.io/spring-kafka/reference/kafka/annotation-error-handling.html#retrying-batch-eh that is...

@chickenchickenlove , would you mind elaborating on your conclusion about `ContainerPausingBackOffHandler`? Thanks

Just tried to crack it someway by propagating a `BackOffHandler` from the `DefaultErrorHandler` down to the `ErrorHandlingUtils.retryBatch`: ``` while (Boolean.TRUE.equals(retryable) && nextBackOff != BackOffExecution.STOP) { backOffHandler.onNextBackOff(container, thrownException, nextBackOff); if (!container.isRunning())...

I'm not sure what you mean with a `BatchListenerFailedException` analogy, but I'd be happy to review some contribution for better understanding how this request could be fixed. Thanks