vrajat commented on PR #12157: URL: https://github.com/apache/pinot/pull/12157#issuecomment-1891440840
> So to make sure I understand: we'll essentially have a gauge per partition per table. And if a new consuming segment is started where the start offset < earliest offset still in Kafka the Gauge will be set to `1` for that partition of that table. > > That gauge will remain at `1` until the next time a consuming segment is created and it will be set to `0` if start offset >= the earliest offset in Kafka. good questions here and we should discuss what is the right thing to do. * I have changed the code to a metric per table. As part of alert clean up, SRE team has requested to remove per partition alerts. So the on-call has to refer to logs to get the lagging partitions. We are also changing existing alerts to be table specific. * Another important question is if ingestion should stop. One worry is that stopping ingestion may make data loss worse as more messages may expire. * If ingestion continues, then when and how should the metric be reset? -- 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]
