Details
-
Bug
-
Resolution: Fixed
-
Major
-
7.6.0
-
None
-
Untriaged
-
0
-
Unknown
Description
We have a cbstat for the number of ActiveStreams per Producer, but that is not exposed to Prometheus (see num_streams in DcpProducer::addStats).
We've hit issues with large numbers of ActiveStreams before. If we have a really large number of ActiveStreams (we've seen ~500K), our "cbstats dcp" call can timeout so it is important to have this metric in Prometheus.
We should expose that value in Prometheus, possibly aggregated by connection type, like the other DCP stats.
Attachments
Issue Links
- relates to
-
MB-58742 Avg. ingestion rate of 999 collections drops from 195K items/s to 57K items/s on build 7.6.0-1516
- Closed