Details
-
Improvement
-
Status: Done
-
Major
-
Resolution: Done
-
None
-
None
Description
I was troubleshooting an issue that a user was running into with the Profiler. The logs made it appear that a second batch of messages were being sent after a period of time, when none actually were.
We found that the second batch was actually just Storm timing out the first batch of messages and re-sending them. This would have been simple to uncover if we had the Kafka topic, partition, and offset details of each message logged at a DEBUG level.
I made those code change while debugging the issue. I thought this might be generally useful to roll into Apache.
Attachments
Issue Links
- links to