gianm commented on PR #18028: URL: https://github.com/apache/druid/pull/18028#issuecomment-2936703573
> That's a fair point, @gianm . In that case, the current logging done the respective `DataSegmentKiller` impls should suffice. If needed, we can improve upon them in follow up PRs. > > With embedded kill tasks, the only drawback would be that these segment delete messages would flood the Overlord logs. But I guess operators can always filter those out based on the logger name. > > Another option could be to direct the task logs to a different log file, same as regular task logs and back them up on deep storage. Overlord would then just log a summary of the task and not all the details. > > Let me know what are your thoughts. I'm not really worried about there being too many logs. I think of this log message as part of the segment lifecycle: first a segment is allocated (if in append mode), then published (always), then marked unused (if no longer needed), then deleted (if killed). So these kill logs shouldn't be "flooding", in terms of volume, any more than logs related to allocation and publish are flooding. I think we already have at least one log message for each of those actions. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
