[09:56:57] <dcausse>	 lunch
[11:55:21] <atsukoito>	 hii!
[12:44:09] <dcausse>	 o/
[12:44:49] <ebernhardson>	 \o
[13:09:37] <inflatador>	 <o/
[14:15:20] <dcausse>	 tested grill-with-docs and was pretty amazed...
[14:16:31] <ebernhardson>	 yea it's a nifty thing!
[14:17:30] <dcausse>	 this was I got after (4hours?) https://gitlab.wikimedia.org/dcausse/cirrus-sup-diagram, diagram is still messy but well.. when you have plenty of links what can you do...
[14:20:01] <ebernhardson>	 pretty cool, did it use tooling for the SVG or it just writes the SVG directly?
[14:20:53] <dcausse>	 I asked it to give me options for diagram formats and chose d2, you have to install d2 but then it writes the d2 source and you convert to svg using d2
[14:20:56] <atsukoito>	 dcausse, ebernhardson: what do you think, is it okay way to split the traffic between production and test clusters? https://gerrit.wikimedia.org/r/1301373
[14:21:13] <ebernhardson>	 it does seem to have captured everything i can think of off the top of my head for how SUP connects to the wider systems
[14:21:35] <dcausse>	 atsukoito: looking
[14:22:00] <ebernhardson>	 atsukoito: hmm, probably ok. Especially since it looks like only testwiki
[14:23:11] <dcausse>	 yes, seems fine to pull a var from InitialSettings in CommonsSettings
[14:23:36] <atsukoito>	 ebernhardson: all wikis will go to `opensearch-ttmserver`, and the testwiki will go to `opensearch-ttmserver-test`
[14:23:48] <dcausse>	 makes sense
[14:24:03] <atsukoito>	 (for now all wikis are going to `opensearch-ttmserver-test`)
[14:24:07] <dcausse>	 did you add opensearch-ttmserver in ProductionServices already?
[14:25:14] <atsukoito>	 dcausse: no, i'll add it now. wanted to check if this way of using InitialSettings is alright (thats why the diff is in wip).
[14:25:40] <dcausse>	 sounds good to me
[14:29:51] <atsukoito>	 i'll test and update the diff then, thanks!
[14:32:26] <ebernhardson>	 read through the sup .md, the prose is actually quite reasonable and covers good topics
[14:34:09] <dcausse>	 yes... the grill thing made really good points and went deeply in asking clarification questions, esp. on fuzzy concepts like active/active double compute/single compute
[14:34:38] <dcausse>	 also found a mistake I made stating that we join on rev_id, we actually join on page_id
[14:36:40] <ebernhardson>	 yea i found the same when putting together planning for first class redirects.  I suppose i'll have to learn more about writing useful skills
[14:37:05] <ebernhardson>	 but even over multiple sessions, focusing on different areas, it had questions i had to think about to answer
[14:40:40] <dcausse>	 the downside is that I spent most of my iterations on the diagram readability, e.g. move X out of the box Y, add bullet points to Z... well probably where I'd have spent most the time doing this myself
[14:45:15] <ebernhardson>	 dcausse: yea same. For the redirects patch, the update one has kinda ballooned to +1800, was thinking to split it? It's at least a few separate things (job handling, saneitizer, api) that could be their own patches, and it's all behind a flag so it shuldn't hurt merging separately
[14:45:21] <ebernhardson>	 just thinking it's gotten pretty excessive to review
[14:46:05] <dcausse>	 yes... for code I could see it going wild pretty quickly...
[15:25:21] <atsukoito>	 dcausse: rollout on monday morning? https://gerrit.wikimedia.org/r/c/operations/mediawiki-config/+/1301373
[15:25:50] <dcausse>	 atsukoito: sure!
[15:31:07] <inflatador>	 ^^ that diagram repo is really cool . I wonder if we could work some of the README.md into our wikitech docs
[15:33:34] <dcausse>	 inflatador: funny thing is that it used some wikitech entries to write that :)
[15:49:32] <inflatador>	 LOL
[16:33:09] <atsukoito>	 see you on Monday, have a good weekend
[16:33:21] <cdanis>	 have a good weekend :)
[16:34:26] <dcausse>	 going offline too, have a nice week-end
[16:51:35] <inflatador>	 .o/