Closed ShahabT closed 1 year ago
Can you add a bit more to the "why", as in specifically how users may use it?
Can you add a bit more to the "why", as in specifically how users may use it?
Great point, @cretz. This is a warm up task for me and I do not have all the context about how users benefit from this. @dnr, could you explain the use case please? (I'll add your reply to the "Why" section)
It's basically just for visibility and completeness. Buffer overruns are already exposed in a metric, but we have similar fields in ScheduleInfo for number of runs skipped, etc. so I think it should go there too.
Buffer size is also interesting to monitor, but it doesn't make sense as a metric, it's more useful as a per-schedule thing to check with DescribeSchedule. I could imagine:
Makes sense. I may request a bit more detail into what the "buffer" is on the proto comment. For example, I don't know whether buffer is used for future calendar/interval or just backlog of backfill or otherwise pending actions from the past.
What changed? Adding two fields to ScheduleInfo proto
Why? This is for visibility and completeness. Buffer overruns are already exposed in a metric, but we have similar fields in ScheduleInfo for number of runs skipped, etc. so I think it should go there too.
Buffer size is also interesting to monitor, but it doesn't make sense as a metric, it's more useful as a per-schedule thing to check with DescribeSchedule. I could imagine:
How did you test it? No test
Potential risks None
Is hotfix candidate? No