-
Notifications
You must be signed in to change notification settings - Fork 336
Closed
Description
We use the process_batch.consumer instrumentation to monitor consumer offset health (e.g. growing offset lag, constant offset lag, very large offset lag, and offset progression). We also have some consumers that use each_message, but for those I'm not sure how to monitor offset lag.
There is an offset lag in process_message.consumer but I don't want to call monitoring for each message, plus it is tricky to monitor offset progression (meaning when I receive a batch in one partition its offsets should start after the previous batch from that partition).
What do you think about instrumenting the batch fetch for each_message as well?
Metadata
Metadata
Assignees
Labels
No labels