[10:26:36] <dcausse>	 hm... extensions/CirrusSearch/maintenance/UpdateSuggesterIndex.php does not work with mwscript-k8s... failing with Http.php: Couldn't connect to host, Elasticsearch down?
[10:27:04] <dcausse>	 testing another script like DumpIndex it's able to connect to elastic...
[11:15:14] <dcausse>	 lunch
[13:13:58] <inflatador>	 <o/
[14:02:27] <dcausse>	 o/
[14:08:03] <pfischer>	 o/
[14:32:10] <inflatador>	 getting a WDQS maxlag alert...checking
[14:32:46] <ebernhardson>	 \o
[14:33:40] <inflatador>	 <o/
[14:35:46] <inflatador>	 https://grafana.wikimedia.org/goto/xXkmJ5WHR?orgId=1 looks like we had a spike, but wdqs1015 seems to be the only one with lag > 5m . Will keep an eye out
[14:35:55] <dcausse>	 o/
[14:38:40] <dcausse>	 seems to be load related, wdqs1015 went above a cpu load of 60
[14:39:16] <dcausse>	 the parallel query_count also increased during that period
[14:39:46] <inflatador>	 Y, 1013 also had ~8m of lag before dropping back down
[14:39:56] <inflatador>	 Hopefully our RAG friends aren't back ;)
[14:41:49] <dcausse>	 yes I hope so
[14:46:00] <ebernhardson>	 hmm, we're going to just miss the 1.43 release with opensearch :(
[14:51:04] <dcausse>	 ?
[14:51:13] <dcausse>	 ah the MW release
[14:51:23] <ebernhardson>	 yea, they are starting branches now
[14:52:22] <dcausse>	 1.43 is going to be LTS :/
[14:53:59] <ebernhardson>	 ouch, i can't see any way we make it into the release with opensearch though, just too tight
[14:55:58] <dcausse>	 yes... way too short, we could perhaps provide a "preview" for 1.43 with a separate branch as soon as we're getting something usable, hopefully that's still compatible with 1.43
[14:56:20] <ebernhardson>	 i mean cindy passes....maybe we call it good enough :P
[14:56:24] <dcausse>	 :)
[16:03:40] <ryankemper>	 ebernhardson: not sure if you had my same confusion, but it was eu that had DST not us
[16:03:51] <ebernhardson>	 ahh!
[17:01:39] <inflatador>	 errand/lunch, back in ~90
[17:05:28] * ebernhardson should get a better understanding of how sonarqube is configured
[17:39:31] <inflatador>	 back
[18:14:06] <inflatador>	 lunch, back in ~30
[18:27:16] <dcausse>	 dinner
[18:36:19] <inflatador>	 back
[19:45:53] <ebernhardson>	 initial numbers in for virtual pageviews. On a day with 450k fulltext sessions 14k of them did a virtual pv (~3%).  Curiously of those 14k sessions 6k abandoned (40%), which is ~2.4% of all abandoned sessions
[19:50:00] <inflatador>	 Is the "virtual pageview" definition still what's shown on this page? https://wikitech.wikimedia.org/wiki/Data_Platform/Data_Lake/Traffic/Virtualpageview_hourly
[19:52:59] <ebernhardson>	 inflatador: yes, should be
[19:53:16] <ebernhardson>	 essentially users hovering a link and getting a popup summary of the article instead of clicking through
[19:54:11] <ebernhardson>	 i suppose my theory on why 40% of virtual pageview sessions abandon would be that those attracted enough by the link simply click it. Users that are tentative and hover their mouse over the link without clicking are more likely to abandon
[19:55:12] <ebernhardson>	 or alternate theory, they were satisfied by the virtual page view and didn't need to click
[20:08:17] <inflatador>	 yeah, that seems reasonable
[20:23:13] <ebernhardson>	 nformation about virtual page views.
[20:23:15] <ebernhardson>	 nformation about virtual page views.
[20:23:44] <inflatador>	 ebernhardson question on the indexing dashboard ( https://grafana.wikimedia.org/goto/WNzeGpWNR?orgId=1 ). Does it make sense to have different update panels for cloudelastic, codfw, and eqiad? Or should we combine?
[20:24:15] <ebernhardson>	 nformation about virtual page views.
[20:25:32] <ebernhardson>	 doh, not sure why i pasted that three times :P
[20:25:36] <ebernhardson>	 inflatador: hmm
[20:26:55] <ebernhardson>	 inflatador: I assume you mean the ones like `CirrusSearch updates - eqiad`? Sadly those are completely wrong now, or at least mis-represent what is going on
[20:28:32] <ebernhardson>	 perhaps they could simply be deleted from the dashboard. Those are the % for the remaining updates like weighted_tag resets or the deleted page archive.
[20:29:31] <inflatador>	 ebernhardson got it. I assume there are similar metrics for the new SUP and we're already using them in a dashboard?
[20:33:52] <ebernhardson>	 inflatador: yea the same data is in the `Action Results` parts of the sup dashboard
[20:36:16] <inflatador>	 ACK, will delete then
[20:46:23] <gehel>	 ebernhardson, dcausse: would you have time to help Dan on T377594?
[20:46:24] <stashbot>	 T377594: Fix Dumps - errors exporting good revisions - https://phabricator.wikimedia.org/T377594
[21:01:27] <ebernhardson>	 gehel: i gave it a read through, i can probably give it a look through but not entirely sure where to start.
[21:08:40] <gehel>	 ebernhardson: can you ping Dan to see where he needs help ?
[21:09:44] <ebernhardson>	 sure, i left a message on slack
[21:11:33] <gehel>	 Thanks ! Olja is asking me to put that on high priority, so whatever you can do to help !