[08:25:26] <dcausse>	 Trey314159: responded, the results they want are at pos 11 and 12 so it's more a ranking issue imo, I don't think anything changed it's just that all other suggestions are matching the "plain" suggestion field and the expected results are matching the icu_folded one. Unless something changed regarding ς and σ which are rightfully (I think) considered identical (most likely thanks to the
[08:25:28] <dcausse>	 icu_normalizer)
[10:39:17] <dcausse>	 lunch
[13:16:51] <inflatador>	 <o/
[13:31:59] <dcausse>	 o/
[14:16:51] <ebernhardson>	 something came up, will be ~30 late for meeting
[14:22:46] <cormacparle>	 hi folks - we have this wish in for CommTech, and I suspect most of what this person is asking for is already available. Any thoughts/comments? https://meta.wikimedia.org/wiki/Community_Wishlist/Wishes/Improve_search_facility/en 
[15:02:32] <dcausse>	 cormacparle: I think this needs more details, I get they want a sanitized version of the source text?
[15:10:20] <gehel>	 not feeling great, I'll end my day early. See you tomorrow !
[15:13:08] <inflatador>	 ;/ feel better soon!
[15:40:45] <ebernhardson>	 ls
[15:58:33] <inflatador>	 workout, back in ~40
[16:49:21] <inflatador>	 back
[16:57:17] <dcausse>	 dinner
[17:29:14] <Trey314159>	 I think I caught whatever is going around the office. I'm going to take a half day.
[17:30:50] <inflatador>	 m( _  _ )m
[17:46:32] <inflatador>	 ryankemper got a small CR for allowing swap on the stat hosts if you have time to look: https://gerrit.wikimedia.org/r/c/operations/puppet/+/1080769
[17:47:03] <inflatador>	 lunch
[18:13:32] * ebernhardson is equally not having luck with the analysis-hebrew gradle plugin build...
[18:42:53] * ebernhardson sometimes wonders how wide the docker developers screens are...`docker ps` is is *wide*
[18:43:26] <ebernhardson>	 also, opensearch doesn't like my plugin build. jarhell :(
[18:56:15] <inflatador>	 (╯°□°）╯︵ ┻━┻
[20:18:09] <ebernhardson>	 cindy is getting happier: 38 passed, 5 failed, 1 skipped, 44 total (100% completed) in 00:11:54
[20:18:25] <ebernhardson>	 (test files, it doesn't detail how many fail in each file)
[20:22:56] <inflatador>	 I'm probably missing something obvious, but what is the prometheus replacement for the graphite metric 'MediaWiki.CirrusSearch.$cirrus_group.requestTimeMs.*.sample_rate' (ref https://grafana.wikimedia.org/goto/PZN4NhiNg?orgId=1 ) ? Was looking at `mediawiki_CirrusSearch_request_time_seconds_sum` in grafana explorer but I can't make the numbers line up
[20:43:54] <inflatador>	 also, we don't care about https://grafana.wikimedia.org/goto/aWs5vhmNg?orgId=1 , do we? Looks like the "job abandonment rates" are tied to the old pipeline?
[21:02:43] <inflatador>	 OK, I think I got the QPS stuff sorted...compare new ( https://grafana.wikimedia.org/goto/QQWoOhiHg?orgId=1 ) to old ( https://grafana.wikimedia.org/goto/7AXod2mNg?orgId=1 )
[21:05:51] <inflatador>	 there's a lot more query types on the new, should we filter some of 'em out?
[21:16:20] <ebernhardson>	 hmm,
[21:16:53] <ebernhardson>	 inflatador: requestTimeMs became request_time_seconds
[21:17:08] <ebernhardson>	 so you should have the right one
[21:18:16] <ebernhardson>	 inflatador: as to how many, hard to say. I'm fine having more things there, the way i use that graph is to filter it to the one thing i'm currently curious about
[21:19:28] * ebernhardson separately can't decide what the right thing is to do for hebmorph on the new lucene version
[21:22:25] <inflatador>	 ebernhardson ACK, thanks for taking a look
[21:22:53] <inflatador>	 and this is what immediately comes to mind when a Texan sees "heb" ;P https://www.heb.com/
[21:23:19] <ebernhardson>	 :P
[21:23:46] <ebernhardson>	 it's complicated by hebmorph being AGPL
[21:34:25] <inflatador>	 Is it because we'd have to mix different code licenses? Or something specific to AGPL?