0:00 AM
ephemer0l has quit
0:00 AM
flamingspinach has quit
0:11 AM
D4RK-PH0ENiX has quit
0:12 AM
prabal
Oh it's alright :)
0:29 AM
milkii_ is now known as milkii
0:34 AM
D4RK-PH0ENiX joined the channel
0:51 AM
tykling
jeg /Win 32
0:52 AM
CatQuest
hi tykling
1:00 AM
shivam-kapila has quit
1:56 AM
ephemer0l joined the channel
2:10 AM
Major_Lurker has quit
2:26 AM
c1e0 joined the channel
2:35 AM
c1e0_ joined the channel
2:37 AM
c1e0 has quit
3:21 AM
shivam-kapila joined the channel
3:26 AM
c1e0 joined the channel
3:29 AM
c1e0_ has quit
3:37 AM
adhawkins has quit
3:39 AM
adhawkins joined the channel
5:51 AM
sumedh joined the channel
6:10 AM
abhishekpanwar joined the channel
6:11 AM
legoktm has quit
6:12 AM
legoktm joined the channel
6:48 AM
prabal has quit
7:11 AM
shivam-kapila has quit
7:45 AM
abhishekpanwar has quit
7:56 AM
BrainzGit
7:56 AM
BrainzBot
8:00 AM
reosarevok
8:00 AM
(and that one above too)
8:01 AM
yvanzo
I fixed the automation rule, you should not worry about labeling, unlabeling, relabeling, ... PRs.
8:01 AM
reosarevok
Thanks!
8:02 AM
yvanzo
It is only triggered on PR opened, closed, marked as ready for review, and edited (title or description).
8:10 AM
reosarevok
8:19 AM
BrainzGit
8:19 AM
BrainzBot
8:24 AM
reosarevok
yvanzo: also please try to update MusicBrainzServerReleaseProcess.md when you have some time to drop fabric :)
8:30 AM
c1e0_ joined the channel
8:31 AM
yvanzo
reosarevok: thanks for volunteering :)
8:32 AM
8:33 AM
c1e0 has quit
8:33 AM
c1e0 joined the channel
8:35 AM
c1e0_ has quit
8:36 AM
reosarevok
Let me see
8:39 AM
Seems good - at least the list it generated seems legit
8:39 AM
bitmap: ^ maybe we can also merge that today
8:39 AM
yvanzo
After all, I brewed this on your demand :)
8:41 AM
reosarevok: please make sure it works for you to update beta.mb.o at least
8:43 AM
reosarevok
I have the config already though now :D
8:44 AM
yvanzo
Is it the same?
8:45 AM
reosarevok
More or less
8:45 AM
I guess I can replace it by the autogenerated one
8:45 AM
And test further
8:45 AM
Then I can actually update syswiki while at it
8:46 AM
yvanzo
Nice, the scripts replacing fabfile should indeed be tested by people set to use it at some point.
8:47 AM
It works for me at least (updated test/beta/prod using it) but we don’t have the same setup.
8:48 AM
reosarevok
Hopefully we can merge a few more PRs today and then I can try to release beta and document it in the evening or tomorrow morning
8:48 AM
sumedh has quit
8:57 AM
sumedh joined the channel
9:00 AM
yvanzo
reosarevok: no worry, you can update beta now if you want to, that doesn't prevent us from updating it afterwards.
9:08 AM
reosarevok
I know, but that reduces the incentive to review more stuff :p
9:11 AM
And stuff has been jumping from milestone to milestone that I really, really would like to actually get merged for once...
9:11 AM
(not only my stuff, fwiw)
9:12 AM
9:14 AM
sumedh has quit
9:24 AM
yvanzo
how does opening for beta test sooner reduce anything? ^^
9:43 AM
niceplace has quit
9:45 AM
niceplace joined the channel
10:03 AM
c1e0_ joined the channel
10:05 AM
Gazooo has quit
10:05 AM
c1e0 has quit
10:07 AM
Gazooo joined the channel
10:08 AM
abhishekpanwar joined the channel
10:14 AM
ruaok
moooin!
10:15 AM
zas: how hard would it be to setup your mb_stats for LB? this is to address the ongoing rabbit mq messages for the LB queue.
10:16 AM
zas
this is simple: it exists already ;)
10:16 AM
ruaok
> For what it's worth, this is literally a spammer reporting someone for downvoting their spam. That's a first.
10:16 AM
zas
10:16 AM
ruaok
lol.
10:16 AM
zas: not that one. I want to know top ips hitting the service.
10:17 AM
zas
ah ;) ok
10:17 AM
let me check
10:17 AM
ruaok
oh, but that is interesting too. it confirms what I already deduced from another view.
10:17 AM
large bursts of data coming in, but all duplicate. all removed, nothing written to DB.
10:17 AM
pointless.
10:20 AM
zas
ruaok: there are multiple nginx log files for LB which one(s) should be parsed?
10:20 AM
ruaok
listenbrainz-web-production is the container that I wish to monitor.
10:22 AM
zas
so it should be prod.listenbrainz-web.access.log
10:22 AM
ruaok
sounds like it.
10:22 AM
zas
there's also a prod.listenbrainz-follow-dispatcher.access.log
10:25 AM
ruaok
yes, not that one.
10:26 AM
yvanzo
In your experience, would using a msg_queue table in PostgreSQL in place of RabbitMQ exchange cause performance issues?
10:27 AM
That is about queuing MB search index messages. I think it does but didn’t have to deal with previous impl.
10:28 AM
ruaok
yvanzo: is that directed at me?
10:29 AM
because I don't really have an answer for you.
10:29 AM
CatQuest
hommik reosarevok
10:29 AM
shivam-kapila joined the channel
10:30 AM
how is everyone? elsa hasn't blow you all away?
10:30 AM
yvanzo
ruaok: to anyone having an experience with the current and previous design.
10:30 AM
ruaok
k. right now it uses rabbitmq, no?
10:30 AM
CatQuest
or drownd you ...
10:31 AM
yvanzo
Yes and I guess it has been chosen for performances.
10:31 AM
CatQuest
10:32 AM
ruaok
yvanzo: yes and it is really quite good at what it does.
10:32 AM
yvanzo
But deploying pg_amqp extension can not work with managed databases (for mirrors only), so using a PG table can be an alternative.
10:32 AM
ruaok
"managed databases (for mirrors only)" can you elaborate on this, please?
10:32 AM
yvanzo
AWS and such I guess
10:33 AM
Online services don't allow for any PG extension.
10:33 AM
ruaok
ah, i see. ok, that provides more context around this.
10:33 AM
zas
10:33 AM
yvanzo
10:34 AM
I think we had a conversation before this PR but I cannot find it again.
10:34 AM
ruaok
zas: <3 <3 <3
10:35 AM
zas
ruaok: stats based on yesterday's log (rotated around 7am)
10:35 AM
ruaok
and now I realize that I need to ask one more thing of you. the API container as well.
10:35 AM
let me look which one that is.
10:37 AM
huh. no, its in the same container. does your stuff also scan for POST methods, zas?
10:38 AM
yvanzo
Anyway, thank you, that already answered my main question.
10:38 AM
zas
not sure, I re-used the parser made for MB website, let me check
10:38 AM
ruaok
yvanzo: by the looks of it that PR doesn't change us away from rabbitmq, right?
10:41 AM
reosarevok
10:41 AM
(since the validation is only negative, but there's no validation for a specific entity type over others)
10:41 AM
(which is why I didn't originally add it, it didn't seem to do much)
10:44 AM
yvanzo
ruaok: if I read it correctly, it does insert an intermediary PG queueing table in front of MQ exchange.
10:45 AM
ruaok
ok, that makes sense now. have they tried it in a production env?
10:45 AM
yvanzo
That might be even more reliable than the current method (I don’t know what pg_amqp is doing when MQ server is down), but will likely impact performances for us.
10:45 AM
yes, IIRC, they use it already for mirroring
10:46 AM
which is nice for a dedicated mirror, but not for us.
10:46 AM
ruaok
maybe just wait and see how it goes for them and then we have an option if we need an alternative?
10:46 AM
zas
ruaok: POST aren't excluded, but you only have few (top IP did only 25 POST, most only few) so they don't really appear
10:46 AM
yvanzo
reosarevok: yes, I didn’t test anything here :)
10:47 AM
reosarevok
yvanzo: is there a point adding [ 'artist', 'event', 'label', 'place', 'series' ] as a "validate it's allowed in general" option?
10:47 AM
ruaok
zas: something is not adding up. there should be TONS of posts. maybe they are hitting the beta container?
10:50 AM
yvanzo
reosarevok: I would say yes, you wouldn't want a bug in 'validate' to block every legitimate url either
10:50 AM
reosarevok
Ok
10:51 AM
sumedh joined the channel
11:02 AM
abhishekpanwar has quit
11:07 AM
yvanzo
ruaok: thanks, replied.
11:22 AM
BrainzGit
11:22 AM
11:23 AM
11:23 AM
11:23 AM
11:26 AM
Shoaib_ joined the channel
11:52 AM
sumedh has quit
11:53 AM
D4RK-PH0ENiX has quit
11:56 AM
sumedh joined the channel
11:59 AM
Gore joined the channel
12:03 PM
shoaib joined the channel
12:03 PM
D4RK-PH0ENiX joined the channel