[07:43:20] <dcausse>	 pfischer: if there's no emergency in getting this feature out we should let the the train do its job imo
[07:44:01] <dcausse>	 we tend to backport only bugfixes
[08:39:55] <dcausse>	 looking at the last 14 days, "Search is currently too busy" are coming only from commons (SpecialMediaSearch to be precise)
[08:46:22] <dcausse>	 and just realized that MediaSearch is also available on some other wikis (ptwikinews) https://pt.wikinews.org/wiki/Especial:MediaSearch?type=page
[09:46:20] <godog>	 hi folks, I've opened https://phabricator.wikimedia.org/T374916 re: categories lag/ping alerts, I'm not quite sure if I got the project tags right though, please adjust at will
[10:00:35] <dcausse>	 godog: thanks!
[10:06:24] <dcausse>	 lunch
[10:42:22] <godog>	 dcausse: sure np!
[13:48:52] <ebernhardson>	 \o
[13:48:58] <inflatador>	 <o/
[13:55:23] <dcausse>	 o/
[14:19:01] <gehel>	 godog: thanks! I'm adding Data Platform SRE as well
[14:57:07] <ryankemper>	 Is it normal to see triple counts going down? I wonder if there's some type of purge happening? https://grafana.wikimedia.org/d/000000489/wikidata-query-service?orgId=1&var-cluster_name=wdqs&from=1724811131811&to=1726584934413
[14:58:34] <ryankemper>	 Steady decrease over last ~6 days. Pattern holds for wdqs-main as well, although not wdqs-scholarly. Which probably lines up with some sort of non-scholarly data being purged
[15:01:08] <Trey314159>	 Could it be related to the mul language code? Merging a bunch of identical labels across languages into one mul label would technically decrease the number of triples, but I don't know if it would be as many as the graph is showing.. though I suppose "millions" is plausible.
[15:02:35] <ebernhardson>	 i imagine millions is likely, i don't think they would go through the trouble if it wasn't
[15:06:04] <dcausse>	 ryankemper: there are a lot of discussions regarding the limitation of the infra since we announced the split (https://www.wikidata.org/wiki/Wikidata:Project_chat#Mass-import_policy) so I would not be surprised that the community started some cleanups and/or started to populate the mul language as Trey suggest
[15:06:37] <dcausse>	 fun fact is that the decrease is only visible on the main subgraph :)
[15:16:20] <dcausse>	 errand
[15:26:28] <godog>	 gehel: sweet, thank you!
[15:39:08] <ebernhardson>	 hmm, naming is hard...not a fan of SiteMatrixPrivateWikiClassifier, but I don't currently have better ideas :P
[15:58:59] <inflatador>	 workout, back in ~40
[16:54:06] <inflatador>	 back
[17:19:28] * ebernhardson mutters at package cycles...never obvious how to fix :P
[17:29:04] <tax>	 The triple count is reducing. The databases are less overloaded. Nature is healing.
[17:34:34] <inflatador>	 lunch, back in ~40
[17:35:25] <ryankemper>	 tax: :P
[18:30:25] <ryankemper>	 inflatador: running 5' late
[18:30:50] <inflatador>	 ryankemper ACK, np
[18:40:01] <inflatador>	 ebernhardson I'm in pairing if interested. Nothing interesting, just looking at T374916
[18:40:02] <stashbot>	 T374916: Port Categories lag / ping checks to Prometheus/Alertmanager - https://phabricator.wikimedia.org/T374916
[18:40:19] <ebernhardson>	 inflatador: oh, i totally missed that while debugging something. Sec
[21:13:42] <inflatador>	 ryankemper +1'd https://gerrit.wikimedia.org/r/c/operations/alerts/+/1073533 but my OCD forced me to update commit msg ;P
[22:48:31] * ebernhardson realizes after writing almost everything...that i have no clean way to delineate private-update-stream being required in producer, but optional in consumer...for tomorrow :P