[12:41:46] <hashar>	 jayme: arnaudb: I have updated the Envoy Telemtry dashboard to use Thanos as a datasource and added site=$site to all metrics!
[12:42:08] <hashar>	 https://grafana-rw.wikimedia.org/d/VTCkm29Wz/envoy-telemetry
[12:44:53] <jayme>	 hashar: thanks. I would suggest to also filter everything by the prometheus label (probably "ops") to query less data
[12:45:33] <jayme>	 please also take a look at the variables which should also use the dynamic $site and $prometheus variables for performant lookups
[12:47:20] <jayme>	 hashar: the envoy k8s dashboard does that for example, see https://grafana-rw.wikimedia.org/d/b1jttnFMz/envoy-telemetry-k8s?orgId=1&editview=variables
[13:51:41] <hashar>	 +1 on prometheus=~"ops"
[13:51:41] <hashar>	 and I will revisit the variables
[13:56:23] <jayme>	 <3
[14:03:24] <hashar>	 I wish we could have a common set of labels that get applied to all the metrics instead of having to amend each of them individual
[14:04:00] <hashar>	 eg `prometheus=ops,site=$site,origin=$origin,instance=$instance`
[14:06:04] <jayme>	 you can maybe put all that in a variable
[14:06:40] <hashar>	 I'll think about it
[14:06:46] <jayme>	 but that would probably make it pretty painful to look at the actual queries a panel uses
[14:06:56] <hashar>	 I have added the filter to the variables, that is a good start