Mr_Monkey: ^^ my comment on that PR is correct then.
2021-02-11 04251, 2021
ruaok
alastairp: thanks.
2021-02-11 04238, 2021
Mr_Monkey
Thanks
2021-02-11 04250, 2021
Clint has quit
2021-02-11 04217, 2021
Mr_Monkey
I thought I was sooo clever when I started using a getter setter on the model, but damn did I get that pye in my face !
2021-02-11 04201, 2021
Clint joined the channel
2021-02-11 04238, 2021
ruaok
heh. when I think a piece of code is clever, it is the first hint that I am on the wrong track.
2021-02-11 04201, 2021
alastairp
ah, that sucks that pydantic doesn't serialise properties
2021-02-11 04221, 2021
alastairp
I guess it makes a bit of sense... it exists so that you can define a data model along with types
2021-02-11 04254, 2021
alastairp
and I'm a little unhappy with the fact that we have 2 of these fields (e.g. collaborators/collaborator_ids) it'd be nice to work out how to combine them together
2021-02-11 04219, 2021
alastairp
maybe one option is to have a User object with id/name fields, but I'm not convinced about that either
Mr_Monkey: _lucifer on ^^ -- should I enter a PR to rename that container?
2021-02-11 04211, 2021
ruaok
that has been bugging me too.
2021-02-11 04214, 2021
Mr_Monkey
Oh, that would be grand
2021-02-11 04220, 2021
ruaok
what should we call it?
2021-02-11 04228, 2021
Mr_Monkey
`websockets_server` ?
2021-02-11 04244, 2021
ruaok
yeah, I was going for something similar.
2021-02-11 04217, 2021
Mr_Monkey
or simply `websockets` I guess, in the context of `listenbrainz-websockets-beta` for example.
2021-02-11 04238, 2021
ruaok
💯
2021-02-11 04230, 2021
ruaok
while we're doing this, should we also remove the old follow feature?
2021-02-11 04233, 2021
ruaok
and maybe do another release?
2021-02-11 04214, 2021
Mr_Monkey
In another release. I think it's bad practice to remove a feature without warning or a replacement, even if the feature is a bit wonky.
2021-02-11 04236, 2021
ruaok
ok fair. I'm 1000% positive no one will miss it.
2021-02-11 04258, 2021
Mr_Monkey
So regarding the consul config, what effect does changing the name of a service have? the vhosts don't change in this case, just the consul service name I guess ("beta.listenbrainz-follow-dispatcher").
2021-02-11 04258, 2021
Mr_Monkey
Does it require a restart of the services, and of consul?
this very large single line command starts the container.
2021-02-11 04203, 2021
ruaok
lines 1331-1137 and 1339-13340 set env vars for the container.
2021-02-11 04209, 2021
Mr_Monkey
Also do you want me to open a PR with the required changes in the LB repo (uwsgi files)?
2021-02-11 04213, 2021
outsidecontext
if the old follow feature is the stuff under the follow menu entry, which I never figured out how to use, than I'm all in favour of getting this removed or bury it real deep
2021-02-11 04224, 2021
ruaok
the registrator service that zas is working on right now picks these up and then updates consul accordingly.
2021-02-11 04235, 2021
ruaok
Mr_Monkey: already done.
2021-02-11 04238, 2021
Mr_Monkey
OK.
2021-02-11 04254, 2021
Mr_Monkey
outsidecontext: Yeah, I agree too. We have a replacement in the works thankfully.
2021-02-11 04259, 2021
ruaok
outsidecontext: yeah, which is why I was for a swift removal of it.
Mr_Monkey: care to test to see if the websockets stuff is working?
2021-02-11 04248, 2021
Mr_Monkey
Sure. On test, beta and prod I imagine?
2021-02-11 04257, 2021
ruaok
most importantly prod.
2021-02-11 04204, 2021
ruaok
I see stuff in logs, so that is a good sign
2021-02-11 04213, 2021
zas
Freso: discourse email issue fixed
2021-02-11 04237, 2021
Mr_Monkey
Yep, WS updates working on prod
2021-02-11 04251, 2021
Mr_Monkey
With #1276 merged we'll have a proper playlist collaborating feature. Woo!
2021-02-11 04252, 2021
Mr_Monkey
I'm working on the tests now.
2021-02-11 04205, 2021
ruaok
sweet!
2021-02-11 04223, 2021
ruaok
and with 1276 merged more heisentests will be fixed too.
2021-02-11 04257, 2021
c1e0 has quit
2021-02-11 04200, 2021
Mr_Monkey
Looks like it's not working on beta though. logs of the websockets_beta container are filled with `init script has no clue which service to start. Set env var CONTAINER_NAME!`
2021-02-11 04225, 2021
ruaok looks
2021-02-11 04231, 2021
Mr_Monkey
same for test looks like
2021-02-11 04259, 2021
ruaok
oh yes.
2021-02-11 04206, 2021
ruaok
my dum dum. hang on
2021-02-11 04241, 2021
ruaok
meh, just tagging a container forces a complete rebuild right now. annoying.
2021-02-11 04250, 2021
zas
ruaok, Mr_Monkey : lb websockets stats deployed, since domain didn't change still visible at same place (but file is different) -> http://stats.metabrainz.org/goto/zbrDTFPGk
alastairp: another issue with ws cache, entity_id is optional
2021-02-11 04202, 2021
_lucifer
in case entity id is not supplied, we have nothing to find what to invalidate
2021-02-11 04236, 2021
alastairp
can you give me an example of something being cached without an entity id?
2021-02-11 04243, 2021
alastairp
oh, this is like a "list of all recordings" ?
2021-02-11 04246, 2021
_lucifer
yes
2021-02-11 04255, 2021
_lucifer
list of any 50 reviews
2021-02-11 04205, 2021
alastairp
so, you'll have to invalidate that one each time
2021-02-11 04221, 2021
_lucifer
how about not cache if entity_id is not available?
2021-02-11 04253, 2021
alastairp
unless you check if the cache exists, and check if this item is in the cache, and only expire if that's the case...
2021-02-11 04213, 2021
_lucifer
yeah, but the issue is the key is generated is cache.gen_key('list', entity_id, user_id, sort, limit, offset, language)
2021-02-11 04205, 2021
alastairp
yeah, maybe just not caching it is a good idea
2021-02-11 04228, 2021
_lucifer
main user of the data is mb anyways which supplies the entity_id
2021-02-11 04239, 2021
_lucifer
so should be much of an issue
2021-02-11 04243, 2021
_lucifer
*should not
2021-02-11 04258, 2021
alastairp
I mentioned this a few days ago - we have a lot of caching in place, but I have no idea how often the caches are actually used
2021-02-11 04244, 2021
_lucifer
once the metrics pr is merged, we can try to work something out
2021-02-11 04205, 2021
ruaok
alastairp: what is the status of your telegraf in BU support work going? I think I may want to track the progress of the MBID mapping using telegraf and grafana.
2021-02-11 04224, 2021
alastairp
ruaok: the PR for a counter is ready
2021-02-11 04249, 2021
alastairp
when you and I discussed it last you showed an interest in being able to just post a value
2021-02-11 04221, 2021
alastairp
and I left it there - I needed to have a chat with zas about how to store that in granfa
2021-02-11 04211, 2021
ruaok
+1 to finishing that
2021-02-11 04233, 2021
alastairp
maybe we can have that discussion now?
2021-02-11 04218, 2021
ruaok
sure, but we need zas, no?
2021-02-11 04256, 2021
alastairp
we do. I was just thinking that it's like a beetlejuice thing
