[07:44:54] <dcausse>	 comp indices seem sane except arzwiki which is still too small in eqiad compared to codfw...
[07:45:56] <dcausse>	 rebuilding manually to see...
[08:12:22] <dcausse>	 numbers are sane now... but not sure what happened :/
[10:13:50] <dcausse>	 lunch
[13:14:16] <inflatador>	 <o/
[13:14:28] <dcausse>	 o/
[13:50:57] <dcausse>	 quick errand
[14:27:53] <gehel>	 dcausse: do you have an estimate of the number (%) of queries on WDQS that would be affected by the split? I think there was something in one of you earlier reports
[14:30:33] <dcausse>	 gehel: in the "random" sample we see 8.8% (https://wikitech.wikimedia.org/wiki/Wikidata_Query_Service/WDQS_Graph_Split_Impact_Analysis)
[16:02:30] <dcausse>	 dinner
[16:11:21] <inflatador>	 Per SRE staff mtg, MW metric migration will be part of the Hackathon. See T363584 
[16:11:21] <stashbot>	 T363584: [wip] HT 2024 StatsLib Migration - https://phabricator.wikimedia.org/T363584
[16:32:42] <topranks>	 inflatador: sorry I missed the discussion on the incident due to the missing lvs vlans 
[16:33:01] <topranks>	 we had a turn up call for one of the new circuits in brazil which conflicted I only joined just as you were finishing 
[16:34:10] <topranks>	 btw I'm happy to work on that Icinga check too, getting the list of pooled IPs would be the part I don't know how to do 
[16:35:27] <inflatador>	 topranks no problem at all, I saw your update re: json outputs and it sounds like we should be able to put it together fairly easily
[16:37:52] * ebernhardson mutters at dayforce which isn't smart enough to be able to schedule a vacation with a holiday in the middle
[16:39:46] * inflatador wonders if there's a Night Force
[16:40:00] * topranks thinks you should call Delta Force 
[16:40:04] <ebernhardson>	 only night court
[16:40:30] <inflatador>	 ;P
[16:55:15] <ebernhardson>	 random stat of the day, external search is 33.3% of pageviews from germany, internal search (not inc related articles) is 3.3% of pageviews.
[16:55:39] <ebernhardson>	 on deskto
[17:29:22] * ebernhardson ships SUP in codfw (at ~25% rate, same as cloudelastic first deploy)
[17:47:05] <ebernhardson>	 everything seems plausible
[18:06:23] <inflatador>	 nice
[18:06:30] <inflatador>	 lunch, back in ~40
[18:41:13] <dr0ptp4kt>	 ebernhardson: i added a calendar invitation for tomorrow to step through metrics stuff a bit more. have appreciated the ongoing updates, btw.
[18:42:17] <ebernhardson>	 dr0ptp4kt: i guess that saying updates have been lacking recently :P  Work has moved into stat1008:~ebernhardson/T358345-query.ipynb, superset was simply too awkward to handle. 
[18:44:15] <dr0ptp4kt>	 lol, i mean, i see the irc things, that's cool :)  yeah, i know superset is sometimes less-than-natural for some of these things. a re-runnable notebook is not bad, and i think sometimes quite fine.
[19:02:14] <inflatador>	 sorry, been back
[20:24:07] <ebernhardson>	 shipped update disabling writes from cirrus, so just SUP is writing to elastic for ~25% of updates in codfw
[20:24:21] <inflatador>	 {◕ ◡ ◕}
[21:05:50] <inflatador>	 ryankemper here's the CR for alerting on mwapi errors: https://gerrit.wikimedia.org/r/c/operations/alerts/+/1025453