nandini12396 commented on PR #21108:
URL: https://github.com/apache/kafka/pull/21108#issuecomment-3662579037

   Thanks so much for the review and reply! I've updated the PR with the tests. 
   
   > If the remoteFetchQuotaBytesPerSecond is set to 25 Mbps and there is a 
message with 30 MB size, will the consumer get stuck? The previous behaviour 
was to allow the consumption to continue.
   No. The behavior follows Kafka's standard quota handling pattern used in 
other fetch paths - we allow at least one fetch to proceed even if it exceeds 
the quota, but subsequent requests will be throttled to bring the average back 
within limits.
   
   >  How can Kafka differentiate whether the error is from remote storage or 
there is an error in processing the response?
   You're right that this is an improvement worth discussing. Currently, the 
implementation releases the quota reservation on any error during remote fetch 
processing. In practice, most errors happen after bandwidth usage? We could 
refine this to only release on specific error types. What's your preference?
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to