[06:07:34] <zpapierski>	 I'm getting my second dose this morning CEST, I should be back around noon (also CEST) 
[06:19:25] <elukey>	 zpapierski: be vigilant on 5g activity on your phone afterwards :D
[06:20:42] <zpapierski>	 That would be surprising, considering a lack of 5g capabilities ;D
[06:23:54] <elukey>	 aahahahha
[09:24:02] <gehel>	 errand + lunch
[10:25:16] <zpapierski>	 unfortunately, no miraculous 5G capabilities appeared, but I'm still hopeful
[10:25:40] <zpapierski>	 at least there's a side benefit of being mostly immune to COVID-19
[10:30:40] <zpapierski>	 dcausse: can I run cindy manually, or do I need CI for that?
[10:33:10] <dcausse>	 zpapierski: I recommend cindy :)
[10:33:57] <zpapierski>	 sure, but how to run those tests?
[10:34:25] <dcausse>	 submit your patch to gerrit cindy will pick it up
[10:34:41] <zpapierski>	 ok, so CI it is
[10:34:42] <zpapierski>	 thx
[10:37:42] <dcausse>	 lunch
[11:09:20] <zpapierski>	 yep, cindy says I broke things 
[11:09:44] <zpapierski>	 there's no way to run it locally ?
[11:54:07] <dcausse>	 zpapierski: there is but it's painful so not sure it's worthwhile, there're two README in tests/integration
[11:55:05] <zpapierski>	 ok, I'll test manually, it'll probably be faster that way
[11:58:20] <dcausse>	 you can also login to cirrus-integ.eqiad.wmflabs, and run sudo -u cindy bash -c 'tmux a' to attach the current tmux session
[11:58:46] <zpapierski>	 I'll be able to see more details on the build?
[11:59:02] <dcausse>	 you'll have to dig into mediawiki logs I suppose
[11:59:09] <dcausse>	 it depends what the error is
[11:59:37] <zpapierski>	 I'll do a manual test first, there's probably some typo or something
[12:00:20] <dcausse>	 also check all the CI failures (not only cindy)
[12:12:27] <zpapierski>	 yep
[13:39:21] <mpham>	 Morten is asking on T279105 whether user engagement metrics (full text query session with dwell time > 10s) is only implemented on desktop, or if it's also on mobile. Can anyone confirm?
[13:39:21] <stashbot>	 T279105: Create/revive Search Platform team metrics dashboard - https://phabricator.wikimedia.org/T279105
[13:50:54] <dcausse>	 errand
[14:50:22] <ebernhardson>	 mpham: all metrics we have are desktop only, we dont even ship the instrumentation to mobile
[14:50:33] <ebernhardson>	 (i guess i can write that there :P)
[14:52:02] <ebernhardson>	 also \o :)
[14:58:35] <ebernhardson>	 also meh...gotta reboot to make meet happy
[15:00:01] <mpham>	 thanks! If you don't, i'll add a comment to the ticket
[15:18:27] <gehel>	 ryankemper: thanks for the work on wdqs servers! Good to see this move forward and go back to a more normal / stable situation!
[15:32:22] <addshore>	 gehel: is the sync happening?
[15:48:03] <ebernhardson>	 oh fun, seem to have lost power :S On 4g for the minute but i miss my extra monitor :(
[15:59:37] <ebernhardson>	 meh, "Our preliminary determination is that your outage was caused by scheduled maintenance work."
[16:41:56] <ebernhardson>	 hmm, now i wonder if i should be using a direct git reference in the airflow mjolnir environment. Then we wouldn't need to build the package
[16:54:22] <dcausse>	 ebernhardson: the mjolnir wheel still needs to be updloaded to archiva no?
[16:54:40] <dcausse>	 this part I don't get I think
[16:58:40] <ebernhardson>	 dcausse: yes, we place it in the artifacts/ directory of wikimedia/discovery/analytics and upload it like the other wheels there
[16:58:57] <ebernhardson>	 the main difference is everything else gets to artifacts/ through make_wheels.py, here it has to be separately built and copied
[16:59:28] <ebernhardson>	 that just sounds silly...i'll play with using a direct git reference
[16:59:34] <dcausse>	 it's some tooling in the airflow repo that does the upload?
[16:59:54] <ebernhardson>	 yea, there is bin/upload_wheels.py in the airflow repo
[17:00:03] <dcausse>	 oh I see
[17:07:43] <dcausse>	 dinner
[17:17:29] <ebernhardson>	 huh, so with a direct git reference in requirements.txt the existing make_wheels.py will fetch and build a wheel, skipping needing that in mjolnir. The wheel will have the reported version though, almost need something like java snapshot versioning for this to work painlessly
[17:49:14] <gehel>	 ryankemper: can we delay our meeting by 30' or 1h ?
[17:49:34] * gehel still has some guest at home
[17:49:49] <ryankemper>	 gehel: absolutely, just ping me when you settle on a time!
[17:50:18] <gehel>	 Ack
[18:51:32] <ryankemper>	 After this current round of re-images I just kicked off, there's two codfw hosts left and the streaming-updater eqiad test host (`wdqs1009`), and then we'll be all done with T280382
[18:51:34] <stashbot>	 T280382: WDQS hosts low on /srv disk space - https://phabricator.wikimedia.org/T280382
[19:00:38] <gehel>	 ryankemper: still at least 30'.
[19:01:37] <ryankemper>	 gehel: ack! goes without saying but take your time, I'll be around
[19:42:02] <gehel>	 ryankemper: il ne there in 5'
[21:31:28] <gehel>	 ryankemper: could you do that change to the depool threshold of the wdqs internal cluster before we forget about it?
[21:31:41] <gehel>	 I can review it tomorrow
[21:32:06] <ryankemper>	 gehel: sure will get a patch up and tag you
[21:32:14] <gehel>	 thx!
[22:45:33] <ryankemper>	 gehel: for tomorrow https://gerrit.wikimedia.org/r/c/operations/puppet/+/698069