This repository has been archived by the owner on Apr 26, 2024. It is now read-only.
Change response time histogram bucket sizes #6804
Labels
A-Metrics
metrics, measures, stuff we put in Prometheus
T-Enhancement
New features, changes in functionality, improvements in performance, or user-facing enhancements.
z-feature
(Deprecated Label)
z-p2
(Deprecated Label)
Description:
The
synapse_http_server_response_time_seconds
histogram metric uses the default bucket sizes of.005, .01, .025, .05, .075, .1, .25, .5, .75, 1.0, 2.5, 5.0, 7.5, 10.0, _INF
.Considering that this also contains the sync endpoint which is expected to be in the 30s range quite often (and sometimes more for complex responses) it would be useful to change the bucket sizes to better reflect the actual values.
Another option would be to split out the sync endpoint response time into its own metric.
The text was updated successfully, but these errors were encountered: