Closed luca-medeiros-reef closed 3 months ago
Yeah, I was worried it would add too much load (as it has to query the db too). A quick single-user locust run shows relevant load but is still in an acceptable range for a Prometheus with a 5~30s refresh rate. Old: avg 34ms New: avg 82ms
We could further improve and refresh db migration metrics less often
I'll check what can be done upstream. Given that the issue was raised over a year ago with no response, I doubt they will go further to merge a PR regarding this. (Wonder if this is expected behavior..)
Will look into Prometheus custom-collectors and come up with a PR upstream.
Closes #186
Every metrics call will get the latest migration-related metrics.