[13:13:18] <ebernhardson>	 \o
[14:03:34] <inflatador>	 <o/
[16:04:09] <inflatador>	 workout, back in ~40
[16:58:14] <inflatador>	 back
[18:01:01] <ryankemper>	 lunch
[18:54:15] * ebernhardson will just assume there used to be more things in this list: foreach ( [ 'plain' ] as $analyzer )
[18:55:55] <Trey314159>	 that was my assumption, too!
[18:56:36] <Trey314159>	 (or maybe cut-n-paste from similar code that got pared down but not quite enough)
[19:00:06] <ebernhardson>	 looks like it was added in 2016 by david, and started that way. Could be copy/paste. https://gerrit.wikimedia.org/r/c/mediawiki/extensions/CirrusSearch/+/310359/26/includes/Maintenance/AnalysisConfigBuilder.php@190
[19:02:01] <Trey314159>	 Ya shoulda been an archeologist, digging up the past like that!
[19:02:09] <ebernhardson>	 lol
[19:58:14] * ebernhardson mutters at unmodifiable map
[19:58:23] <ebernhardson>	 i mean i like them, but i want to remove something from the map :P
[20:43:12] <inflatador>	 ryankemper or anyone else , patch up for stat host alerts if y'all have the cycles: https://gerrit.wikimedia.org/r/c/operations/alerts/+/1064436
[20:44:40] <ryankemper>	 looking
[20:45:46] <inflatador>	 cool. I'm alerting on load15 > 100 for 1 minute
[20:46:13] <inflatador>	 here's an example of the last lockup: https://grafana.wikimedia.org/goto/SIODWqCSg?orgId=1
[20:46:58] <inflatador>	 it's a matter of personal opinion, but IMHO alerting on load15 is better since it avoids short-lived spikes
[20:51:41] <ryankemper>	 inflatador: gave my +1. I was trying to decide if we should require more than a minute but since it's 15min load average that's kind of baked in already
[20:57:46] <inflatador>	 ryankemper ACK, that was my thought as well...we can revisit as needed though
[21:23:19] <ebernhardson>	 hmm, I wonder if the saneitizer rerender frequency is too low. It's set to 16, with a loop duration of 14 days.  I think we were intending to set it to 16 weeks and not 16 loops
[21:26:10] <ebernhardson>	 oh nevermind, for some reason it's 16 default in the code, but 8 in the helmfile
[22:04:14] <ryankemper>	 With the exception of wdqs2024 which is having reimage problems that we're still troubleshooting, all the remaining data xfers will be done in an hour or so
[22:05:16] <ryankemper>	 Brian and I are going to work on merging the DNS, conftool-data etc patches after SRE retro tomorrow. That will get us up to step #4 inclusive on the steps listed here: https://wikitech.wikimedia.org/wiki/LVS#Add_a_new_load_balanced_service and hopefully we can get a bit further past that as well
[22:09:09] <ebernhardson>	 \o/