You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This command exposes important information about the current offset and lag per partition per consumer group
Describe the solution you'd like
Expose information from that command + on kafka topics. Specifically:
How many messages in each topic (min, max offset per partition)
Current offset, lag per consumer group
This can then be leveraged to figure out if some service is down. E.g. clickhouse-ingestion lagging indicates plugin-server is down, and it being further behind than topic/partition min indicates data loss!
Describe alternatives you've considered
Document manual commands only.
Additional context
This might be a viable approach for exposing this data.
Took a brief dig - KafkaAdminClient is woefully underpowered for this right now (no way to get min/max offsets, lag etc) but it's possible to do this by extending the library in relatively simple ways. Related: dpkp/kafka-python#2278
This issue hasn't seen activity in two years! If you want to keep it open, post a comment or remove the stale label – otherwise this will be closed in two weeks.
Is your feature request related to a problem?
I've diagnosed multiple outages by leveraging kafka consumer group statistics, e.g. via a command like the following:
This command exposes important information about the current offset and lag per partition per consumer group
Describe the solution you'd like
Expose information from that command + on kafka topics. Specifically:
This can then be leveraged to figure out if some service is down. E.g.
clickhouse-ingestion
lagging indicates plugin-server is down, and it being further behind than topic/partition min indicates data loss!Describe alternatives you've considered
Document manual commands only.
Additional context
This might be a viable approach for exposing this data.
cc @fuziontech, @tiina303, @yakkomajuri and @guidoiaquinti for prioritization.
Thank you for your feature request – we love each and every one!
The text was updated successfully, but these errors were encountered: