[06:34:24] <jayme>	 \i/
[07:59:09] <matthieulec>	 Nice, thanks for the update Amir
[07:59:28] <matthieulec>	 looking forward to see the impact
[11:48:21] <claime>	 https://grafana.wikimedia.org/goto/bfvnegzhgpp8gf?orgId=default
[12:31:44] <claime>	 https://phabricator.wikimedia.org/T366938#12229060 Cool to see the immediate impact
[13:29:32] <Raine>	 pretty!
[13:29:54] <Raine>	 a graph that looks good for once! :D 
[13:50:35] <_joe_>	 that change means that our redis servers are more critical than before; please document it where appropriate
[14:57:58] <claime>	 _joe_: It's a separate group of redis servers
[14:58:36] <claime>	 Amir1: mediawiki can handle one of three going down at a time right?
[14:59:12] <Amir1>	 yes, it can handle it no problem
[14:59:35] <claime>	 If two go down what happens? :P
[14:59:58] <Amir1>	 one third of cases would be slow but there might be issues here and there
[15:00:30] <Amir1>	 on the risk matrix of likelihood vs impact. It's acceptable risk IMHO
[15:04:19] <_joe_>	 claime: yeah my point was we can't reboot two of them at the same time
[15:04:22] <_joe_>	 for instance
[15:04:33] <_joe_>	 and/or if we might want to do failover to the replica if that happens
[15:05:05] <claime>	 There's no replica, they're standalone servers
[15:18:04] <Amir1>	 yeah, as long as it's done one by one, anything is okay
[15:20:04] <Amir1>	 they are basically temu version of redis HA cluster
[15:20:19] <_joe_>	 so much disrespect to temu smh
[15:20:37] <_joe_>	 but yeah, my point was "there is some care to apply when rebooting them"
[15:22:10] <Amir1>	 about to deploy it everywhere (https://gerrit.wikimedia.org/r/c/operations/mediawiki-config/+/1327138) now https://usercontent.irccloud-cdn.com/file/bGMTJRo6/grafik.png
[15:34:59] <Amir1>	 the memory usage has shot up... to 14MB... in total