#metabrainz

/

      • JadedBlueEyes has quit
      • JadedBlueEyes joined the channel
      • Kladky has quit
      • huhridge has quit
      • bitmap[m] has quit
      • pite_ has quit
      • reosarevok[m] joined the channel
      • reosarevok[m]
        yvanzo: around and happy to do more testing as needed :)
      • bitmap: do you know if MBS-13832 is intentional? Is it because at some point pdf thumbnails were not generated? They seem to work just fine now so we should probably support them?
      • BrainzBot
        MBS-13832: Poster thumbnail is not generated if the poster is a PDF https://tickets.metabrainz.org/browse/MBS-13832
      • reosarevok[m]
        yvanzo: nvm, saw the email :)
      • Kladky joined the channel
      • mayhem[m] joined the channel
      • mayhem[m]
        lucifer: https://stats.metabrainz.org/d/OGg5QUCGz/listen... the spotify writer is piling up incoming listens again. should I go take a look at what is going on?
      • lucifer[m] joined the channel
      • lucifer[m]
        mayhem: i think its an issue with the metric. the container seems sane.
      • mayhem[m]
        what is weird about it is that when you restart the container, that metric slowly drops.
      • lucifer[m]
        because this one looks fine.
      • mayhem[m]
        yeah, I agree it looks fine on the surface, but....
      • lucifer[m]
        can you check if the grafana query is correct?
      • mayhem[m]
        ok, I'll put it on my list to investigate. I am hunting down all threads of "wtf" we have in our systems now.
      • lucifer[m]
        mayhem[m]: the logs of the container as well.
      • mayhem[m]
        ok, maybe its just fixing the metric then.
      • lucifer[m]
        great, thanks
      • mayhem[m]
        lucifer: I checked the spotify queue and its empty. who knows what that metric is doing, lol.
      • lucifer[m]
        i'll check the grafana query and the spotify code setting the metric to fix it.
      • monkey[m] joined the channel
      • monkey[m]
        Last time I saw the queue growing and inspected the logs, I didn't see anything wrong. However after restarting the container and looking at logs, it was very obvious that it was suddenly doing a lot more imports. That made me think that during those wtf moment it is importing listens for some users, but silently not importing them for all users, hence the growing queue
      • mayhem[m]
        but I didn't actually see a queue with anything in it when I checked the rabbit mq queues.
      • monkey[m]
        But wouldn't a failing spotify reader fail to create messages to add to the queue?
      • I'm looking at the spotify reader logs now, we're importing for 2-3 users every minute which is definitely not normal
      • mayhem[m]
        where would these listens be stored? redis? not in the container themselves, because the listens survive a restart.
      • monkey[m]: ah yes, there was that. deffo not normal.
      • monkey[m]
        Ah it started fetching a few more users a minute
      • mayhem[m]
        one of the tricky things here is that we're dealing with an unstable API -- spotify
      • monkey[m]
        All I can see in the lgos that is duboius is this error that happens fairly regularly:
      • `ERROR Unexpected error while getting listens: HTTPSConnectionPool(host='api.spotify.com', port=443): Read timed out. (read timeout=5)`
      • Yeqh, thqt
      • lucifer[m]
        <monkey[m]> "I'm looking at the spotify..." <- https://stats.metabrainz.org/d/OGg5QUCGz/listen...
      • it is sort of normal, looking at the graph.
      • except for two weeks, it used to be simialr to this rate.
      • monkey[m]
        Yeah, looks normal.
      • BrainzGit
        [listenbrainz-server] 14amCap1712 merged pull request #3044 (03master…fix-art-escape): LB-1686: Fix escaping in art templates https://github.com/metabrainz/listenbrainz-serv...
      • mayhem[m]
        that whole incoming listens metric graph is utterly busted. all the metrics are rubbish, its hard to see though.
      • it says there are 2.3k incoming listens.
      • BrainzGit
        [listenbrainz-server] 14mayhem opened pull request #3047 (03master…fix-spotify-reader-metrics): Fix spotify reader metrics https://github.com/metabrainz/listenbrainz-serv...
      • mayhem[m]
        monkey: ^^
      • mayhem[m] uploaded an image: (191KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/AmhUzbIBkFTNjrDjYCmwYmZZ/Screenshot_20241125-125537.png >
      • akshaaatt: The LB has become unusable for me since the button bar conflicts with the mobile's navigation buttons. Have you seen this before?
      • BrainzGit
        [listenbrainz-server] 14mayhem merged pull request #3047 (03master…fix-spotify-reader-metrics): Fix spotify reader metrics https://github.com/metabrainz/listenbrainz-serv...
      • SigHunter has quit
      • SigHunter joined the channel
      • [listenbrainz-server] 14amCap1712 merged pull request #3045 (03master…errors-json): Fix 404 errors https://github.com/metabrainz/listenbrainz-serv...
      • [listenbrainz-server] 14MonkeyDo opened pull request #3048 (03master…recs-update-text): Add the StatsExplanationsModal to the created for you page https://github.com/metabrainz/listenbrainz-serv...
      • lucifer[m]
        monkey: mayhem has there been any discussion about this year's year in music so far?
      • monkey[m]
        A little bit at out last team video call. The gist of it is this year we want to avoid adding or changing much in terms of data calculations, and only modify the look of the page
      • (Because YIM is coming very soon)
      • If anything, we could consolidate the addition of genre stats from last year, but I think that would be about it.
      • lucifer[m]
        i see, makes sense
      • akshaaatt[m] joined the channel
      • akshaaatt[m]
        Damn mayhem ! Is that related to an upgrade to android 14?
      • jasje: you last made a beta release. Do you recognise this bug?
      • Let me check in the meantime
      • jasje[m] joined the channel
      • jasje[m]
        mayhem: can you tell the version of your app?
      • akshaaatt: no changes related to this since the app is working fine on my end atm with latest version
      • outsidecontext has quit
      • irimi1 has quit
      • djl has quit
      • ursa-major has quit
      • mebious has quit
      • RetroPunk has quit
      • serra has quit
      • ajhalili2006 has quit
      • ajhalili2006 joined the channel
      • serra joined the channel
      • mebious joined the channel
      • irimi1 joined the channel
      • djl joined the channel
      • RetroPunk joined the channel
      • ursa-major joined the channel
      • outsidecontext joined the channel
      • <akshaaatt[m]> "Damn mayhem ! Is that related to..." <- No such changes in android 14, but 15 yes
      • monkey[m]
        lucifer: Am I reading the LFM loved tracks import code correctly, and we only check if the `mbid` in the LFM loved tracks corresponds to a track_mbid? i.e. we don't check if it could be an existing recording_mbid ?
      • lucifer[m]
        monkey: yes
      • monkey[m]
        Some users noticed that only some of their loved tracks are imported, and looking at the LFM endpoint response, some of the items have track mbids while others have recording mbids, both in thye same field :/
      • OK, thanks, I'll create a ticket with my finding
      • s
      • lucifer[m]
        i see, i'll update it to check both. thanks for the ticket!
      • monkey[m]
        Err. I mean I'll add details to LB-1689
      • BrainzBot
        LB-1689: LastFM loved tracks imports does not import all tracks https://tickets.metabrainz.org/browse/LB-1689
      • reosarevok[m]
        bitmap: I see our new Sentry does not talk to Jira so it no longer has the option to create a new ticket from the error, we should maybe look into bringing that back unless it was causing issues
      • BrainzGit
        [musicbrainz-server] 14reosarevok opened pull request #3415 (03master…MBS-13834): MBS-13834: Load l() in Role::Art https://github.com/metabrainz/musicbrainz-serve...
      • reosarevok[m]
        (found that legit issue in between a lot of broken userscript cruft)
      • BrainzGit
        [musicbrainz-server] 14reosarevok opened pull request #3416 (03master…MBS-13829): MBS-13829: Also support Audiomack "song" for releases https://github.com/metabrainz/musicbrainz-serve...
      • mayhem[m]
        <jasje[m]> "mayhem: can you tell the version..." <- I can't use any of the features of the app, so I can't get to the about box, but the play store says: 2.7.2 updated on Sept 26
      • minimal joined the channel
      • reosarevok[m]
        bitmap: also, any idea if the user is right about issues existing in MBS-13826 ? See my comment, I'm confused
      • BrainzBot
        MBS-13826: JSON-LD values sometimes strings instead of URLs https://tickets.metabrainz.org/browse/MBS-13826
      • jasje[m]
        mayhem: android 15?
      • mayhem[m]
        yep
      • jasje[m]
        Google force enabled a setting which we set to off )
      • Because it suited our app
      • Ill release a fix
      • pite joined the channel
      • bitmap[m] joined the channel
      • bitmap[m]
        <reosarevok[m]> "bitmap: do you know if MBS-13832..." <- I don't think so, I'm guessing it was just never updated after we got pdf thumbnails. it doesn't work for releases either
      • BrainzBot
        MBS-13832: Poster thumbnail is not generated if the poster is a PDF https://tickets.metabrainz.org/browse/MBS-13832
      • TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged and not empty as it is bridged to IRC; see https://musicbrainz.org/doc/ChatBrainz for details | Agenda: Reviews, X-odus (reo), block temp email domains (reo), docker-server-configs (mayhem)
      • reosarevok[m]
        bitmap: also, I see there's some Jade fixes - I checked and it seemed sensible but you know a lot more
      • If it looks good we could beta it today
      • Oh, the pdf thing is because of the index_listing view, isn't it. Does that make it a schema change?
      • Seems likely, annoying
      • bitmap[m]
        <reosarevok[m]> "Oh, the pdf thing is because..." <- it would be a schema change but wouldn't affect replication, so we could deploy it to production in meantime
      • reosarevok[m]
        Hmm. Does the change mean people would get PDFs as fronts in the CAA API when they currently don't? If so I think that's fine but maybe it needs to be notified or something
      • bitmap[m]
        I did find some old messages indicating they were excluded because thumbnails weren't available, but then support was added a couple months later, so I guess just the view wasn't updated
      • BrainzGit
        [musicbrainz-server] 14reosarevok opened pull request #3417 (03master…MBS-13832): MBS-13832: Don't skip PDF files for frontiest check https://github.com/metabrainz/musicbrainz-serve...
      • reosarevok[m]
        I guess it'd be something like that
      • LupinIII
        ...
      • reosarevok[m]
        <BANG>
      • Welcome to yet another MetaBrainz Monday Meeting!
      • We have three (3) mailed in notes today
      • Make that 4, @atj with the last second goal!
      • atj says:
      • '''
      • Last week I worked on fixing some GitHub CI issues with the Solr and ZooKeeper Ansible roles. I was then able to test various changes in my development branches and merge them.
      • I also created a couple of branches in our main Ansible repo, one for resolving some deprecation warnings and another in preparation for the Solr 9.7.0 upgrade. These will be raised as PRs this week.
      • '''
      • outsidecontext says:
      • "Last week I had little time, I'm currently alone with the kids. Also I fell sick on the weekend, but already feeling a bit better today. Managed to do some support on the forums and some code review today, though."
      • derat says:
      • "Not much to report this week. I updated yambs to allow bursting up to 5 requests in 5 seconds, so it should be a bit faster in cases where e.g. both artist and label URLs need to be looked up for Bandcamp releases."
      • And yvanzo says:
      • '''
      • Last week, I kept mainly working on the Solr 9 mirror upgrade, unsuccessfully.
      • It is therefore postponed again until a new date is decided, likely next year.
      • Mirror owners don’t have to change anything in the meantime.
      • I’m taking a week off to keep myself from collapsing again.
      • I won’t connect to the chat but I will check the emails.
      • '''
      • Have a good rest, yvanzo!
      • That is it for mails
      • On my list today: aerozol, ansh, mayhem, akshaaatt, yellowhatpro, lucifer, julian45, MonkeyPython, zas, kellnerd, reosarevok, bitmap, jasje, monkey
      • I don't think aerozol is around yet, so ansh?
      • Oh he is!
      • Go then!
      • aerozol[m]
        Mōrena! Sorry just dragged my zzz’s to the compy now