[10:33:36] <gehel>	 pfischer, Joal: connection issues, trying to connect back
[10:49:56] <dcausse>	 errand + lunch
[11:44:09] <pfischer>	 lunch
[13:59:43] <inflatador>	 <o/
[14:14:42] <inflatador>	 Just merged https://gerrit.wikimedia.org/r/c/operations/puppet/+/974281 so we might get some alerts. Not exactly sure what 'severity:task' means...maybe it will automatically create phab task? Will monitor
[14:32:53] <dcausse>	 o/
[14:44:30] <inflatador>	 lexemes reload finished on wdqs1022 ... I updated https://phabricator.wikimedia.org/T347504 but LMK if there's anything I can do
[14:46:35] <dcausse>	 inflatador: thanks!
[14:49:12] <inflatador>	 np
[15:33:31] <ebernhardson>	 inflatador: it should make a task, i configured that here: https://gerrit.wikimedia.org/r/c/operations/puppet/+/769131/4/modules/alertmanager/templates/alertmanager.yml.erb
[15:33:40] <ebernhardson>	 i don't remember how or why that particular webhook string works :P
[15:46:05] <gehel>	 @team: Have I missed something important in our team update: https://docs.google.com/presentation/d/1Hi3hZCD7BLPRqPPdHvdxShO7iLTIykROf5AJLFnEBkU/edit#slide=id.g29cc179a43e_0_300
[15:47:19] * ebernhardson is shocked to see all the flink bits running with a 4d+ age
[16:12:47] <pfischer>	 ebernhardson: that’s great news! A post-thanks-giving gift. ;-)
[16:15:14] <inflatador>	 ;)
[17:15:27] <inflatador>	 https://gerrit.wikimedia.org/r/c/operations/puppet/+/977704 is up for the LDF endpoint monitoring...previous patch failed due to duplicate resources
[18:14:08] <inflatador>	 lunch, back in ~40
[18:24:05] <ebernhardson>	 Anything i shouldn't forget to include in the elasticsearch deep dive tomorrow? Putting together an outline
[18:30:40] <ebernhardson>	 randomly curious (and probably not particularly useful :P) the developer that made most of the early lsearchd contributions now turns up as an author (out of 20 or 30+) of the llama 2 paper from meta
[19:20:55] <inflatador>	 back
[19:29:17] <dcausse>	 dinner
[19:34:32] <inflatador>	 ebernhardson thanks for the Puppet link. I just merged above patch for LDF monitoring...keeping an eye on the discovery-search board to see if we get any alert tasks
[19:45:27] <ebernhardson>	 getting some lunch
[20:39:25] <ebernhardson>	 back
[21:01:02] <dcausse>	 spent some time on friday exploring the elastic connector to allow plugin a "response handler" (https://github.com/nomoa/flink-connector-elasticsearch/tree/bulk-item-tesponse-handler)
[21:02:17] <dcausse>	 not really tested nor something I'm particularly happy with... tracking the number of individual retries seems particularly tedious and found no ways to plug a side-output at this stage :/ 
[21:03:58] <ebernhardson>	 dcausse: thats still pretty cool! I pondered it briefly but wasn't sure how to go about it.  
[21:04:22] <ebernhardson>	 and yea the way that pending actions is a simple counter seems awkward
[21:04:28] <inflatador>	 nice
[21:09:36] <dcausse>	 I think it'd be more correct to have a side-output for individual retries (retrying in the sink might treat events out-of-order, possibly erasing newer data I think?)
[21:10:35] <dcausse>	 well we still have the super_detect_noop version check so perhaps acceptable... 
[21:20:41] <inflatador>	 Another PR for LDF monitoring if anyone has time to look https://gerrit.wikimedia.org/r/c/operations/puppet/+/977787
[22:32:45] * ebernhardson wonders if the Ltr query still needs a +10000 on it's score. I think we put that there because the model could emit negative numbers and we wanted to make sure -5 was still ranked above the other things