From isabel.peters@mongodb.com:
Our system on Cloud Intel to dynamically adjust the profilerLevel slowms value for customer clusters uses serverStatus.opLatencies to determine what the slowms value should be set to for collecting slowlogs. I have just learned that the serverStatus.opLatencies will remain based off durationMillis rather than workingMillis in 8.0 which would be a problem for the system to accurately determine the right profiling slowms threshold
Would it be possible to add opLatencies based off workingMillis in the serverStatus response as well?
We may want to consider adding a different serverStatus section rather than changing the meaning of opLatencies.
- is depended on by
-
COMPASS-7936 Investigate changes in SERVER-89125: Report workingMillis in serverStatus opLatencies
- Closed
- is duplicated by
-
SERVER-89114 Measure opLatencies with workingMillis
- Closed
-
SERVER-89281 Set waitForTicketDurationMillis in _sumBlockedTimeTotal
- Closed
- related to
-
SERVER-90711 Arithmetic error calculating workingMillis in CurOp
- Closed