[07:23:35] hello folks [07:23:40] created https://github.com/wikimedia/revscoring/pull/521 for the new revscoring version [07:23:46] hopefully this time it will be good :) [07:24:19] once merged, I'll use Docker and python 3.7 to build the wheel, and I'll upload it to Pypi [07:25:37] thanks kevinbazira ! [07:26:16] 🙏🙏🙏 [07:49:17] https://pypi.org/project/revscoring/2.11.4/ [07:51:21] 10ORES, 10Machine-Learning-Team (Active Tasks), 10Patch-For-Review: revscoring feature extraction error for wikitext papes in Wikidata - https://phabricator.wikimedia.org/T302851 (10elukey) Published https://pypi.org/project/revscoring/2.11.4/ from a python 3.7 environment (just to be extra sure). The size o... [08:43:00] 10Machine-Learning-Team, 10Add-Link, 10Growth-Team, 10incubator.wikimedia.org: Integrate the model training and the deployment of "Add a link" to new Wikipedias exiting the Incubator - https://phabricator.wikimedia.org/T308146 (10kostajh) >>! In T308146#7934769, @Trizek-WMF wrote: >>>! In T308146#7924078,... [08:45:38] 10Lift-Wing, 10Machine-Learning-Team (Active Tasks): Load test the Lift Wing cluster - https://phabricator.wikimedia.org/T296173 (10elukey) Interesting discovery - it seems that my previous tests with ab and siege used http/1.0, not 1.1, and the responses from istio where all 426 upgrade required (so not reall... [09:48:09] 10Lift-Wing, 10Machine-Learning-Team (Active Tasks): Load test the Lift Wing cluster - https://phabricator.wikimedia.org/T296173 (10elukey) It is very weird, siege supports HTTP/1.1 but I see the following: ` elukey@deploy1002:~$ siege "https://inference.svc.eqiad.wmnet:30443/v1/models/enwiki-goodfaith:predic... [13:21:41] morning all [13:22:09] I'm going to be grinding silently all day. I have 100 things complete today plus a presentation at applyconf later [13:37:26] morning, good luck for the presentation! [14:01:25] 10Lift-Wing, 10Machine-Learning-Team (Active Tasks): Load test the Lift Wing cluster - https://phabricator.wikimedia.org/T296173 (10elukey) I was able to use `wrk`, very interesting tool installed on deploy1002. We can use lua scripts like the following: ` elukey@deploy1002:~$ cat inference.lua wrk.method =... [14:01:57] reported my findings about load testing a kserve pod in --^ [14:02:21] I think that we still have some bottlenecks here and there, plus we need a score cache :) [14:03:25] 10Machine-Learning-Team, 10Add-Link, 10Growth-Team, 10incubator.wikimedia.org: Integrate the model training and the deployment of "Add a link" to new Wikipedias exiting the Incubator - https://phabricator.wikimedia.org/T308146 (10MGerlach) @kostajh I dont think there is a well-defined minimum. In principle... [14:05:23] going afk for a bit, ttl :) [14:06:30] (sad thing - we'll probably need to rely on change-prop and pre-caching scores to get good performances, replicating bits of ORES in k8s, but with the revscoring models we probably cannot do more) [15:03:03] klausman pinging for the interview [15:03:06] omw