[00:49:50] <Krinkle>	 godog: https://grafana-rw.wikimedia.org/d/IROgrArMk/paint-timing?orgId=1 it seems the painttiming metric has stopped being reported, which I asume is due to the ops->ext switch, but after switching the datasource to Thanos, data still stops.
[00:51:31] <Krinkle>	 (no alert fired for us as we alert on the graphite source through other dashboards still)
[01:04:48] <AaronSchulz>	 Krinkle: do we use "sets" for any metrics? https://github.com/statsd/statsd/blob/master/docs/metric_types.md
[01:08:05] <Krinkle>	 AaronSchulz: not within MediaWiki, no.
[01:08:13] <Krinkle>	 not in our own perf services either
[01:09:11] <Krinkle>	 note that we use a different implementation of both the proxy and the lb than the original Statsd, links at https://wikitech.wikimedia.org/wiki/Statsd
[03:33:28] <AaronSchulz>	 Krinkle: I refactored ping() in https://gerrit.wikimedia.org/r/c/mediawiki/core/+/859651
[08:25:32] <godog>	 Krinkle: I see the data now, I'm assuming you fixed paint-timing ?
[13:31:13] <Krinkle>	 godog: hmm strange. Well I guess it fixes itself
[14:40:18] <godog>	 Krinkle: indeed, I noticed the dashboard is a little slow to load, we can look into turning some/all expression into pre-recording rules to speed things up fwiw