GitLab Outage

(status.gitlab.com)

69 points | by kunley 2 days ago ago

36 comments

  • davb 2 days ago ago

    Coincidentally (?) we’ve seen a huge uptick in failed login attempts to our MFA-protected admin accounts on our self-hosted Gitlab CE instance today. They’re coming from thousands of IPs across hundreds of networks (VPS providers and, apparently, residential proxies).

    • deckar01 2 days ago ago

      There was a relative lull in GitHub bot activity the last few weeks, but they seem to be back today. I assume the promo token flood gates opened for the new models.

      • physicallyIllfr 2 days ago ago

        I dont think bots need bonus/trials. There are lots and lots of ways to get free, or virtually free tokens. Hell half the vibe coded AI chrome extensions send store their openrouter keys client side.

    • TZubiri 2 days ago ago

      It's probably related to the issue published by aikidolabs where leaked per-user gitlab issued emails were being used as a sort of credential.

      https://www.aikido.dev/blog/gitlab-email-push-to-main

  • b0afc375b5 2 days ago ago

    Interesting that the status page was updated immediately. Is this the norm now? I remember back in the day when status pages showed all green even when servers were clearly down for a long time already.

    • stack_framer 2 days ago ago

      It was not updated immediately. I was seeing 503s for a solid 15 minutes while the status page was all green. This is also their second outage today:

      https://status.gitlab.com/pages/history/5b36dc6502d06804c083...

      • nicholasjarnold 2 days ago ago

        Yep, status page was certainly all green when I started noticing it.

    • seabrookmx 2 days ago ago

      As others mentioned, it wasn't that quick.

      It's also not accurate. After nearly a half hour the page shows "git operations" as green but I'm not able to push/pull from my repos with SSH or HTTPS (and haven't been able to the whole time).

    • thesh4d0w 2 days ago ago

      It took at least 5 minutes, closer to 10. I think the outage is so bad, it got pushed fast.

    • kunley a day ago ago

      It was updated at least 5 minutes after I noticed the outage first. Guess it's typical organisation lag

  • delf 2 days ago ago

    You can use GitSocial to mirror to your own bucket for situations like this: `gitsocial mirror https://gitlab.com/owner/repo s3://s3.example.com/mybucket/repo`

    https://gitsocial.org/

    • PhilippGille a day ago ago

      It would be nice to disclose that you're the creator of the tool which you are promoting here.

      • delf a day ago ago

        Will do it going forward. Thank you for the feedback!

  • hk1337 2 days ago ago

    I like what tangled seems like it could be but it's still too new to really use. It'd be nice if it was more of an aggregator for git servers where still access a git server whether tangled is up or not.

    • mitxela 2 days ago ago

      You don't really need one of these systems. You can just host git.

      • hk1337 2 days ago ago

        True, my thought was a site that just makes it easier to see all the various users. Like GitHub but links to users and orgs actually go to each person’s server.

  • theplumber 2 days ago ago

    Maybe Sam or Dario is playing hide and seek again with the agents….

  • lampcord 2 days ago ago

    Another day, another reminder why we self-host our GitLab instance. Cloud convenience always has these kinds of trade-offs.

    • zer0x4d 2 days ago ago

      We self host ours too. 100% uptime ;)

      • gib444 2 days ago ago

        How do you do upgrades?

      • TZubiri 2 days ago ago

        what host are you using that gets 100%uptime? A hyperscaler like AWS?

        • mitxela 2 days ago ago

          Just means it hasn't crashed yet.

        • zer0x4d 2 days ago ago

          I run it on Digital Ocean but AWS should be fine too.

        • firesteelrain 2 days ago ago

          VMWare Cloud Foundation with Dell R7725s

  • nabeards 2 days ago ago

    Just went to push code and found this out.

    • nicholasjarnold 2 days ago ago

      Code review for me... Hit the 503. The status page lagged the event by ~10 mins or so it seems.

      git+ssh also seems affected, as I'm unable to push/pull right now.

  • ChrisArchitect 2 days ago ago
  • smalltorch 2 days ago ago

    Gasp! immediately builds a complete git alternative to never expirence a single second of downtime ever again

    • JonChesterfield 2 days ago ago

      Well, git is already distributed, so you could use _any other computer_ instead of gitlab.com...

      • smalltorch 2 days ago ago

        I'm only kidding, although it could be a good excersise to learn. I think a pure TUI git front end / back end project would be fun to make.

      • gblosser 2 days ago ago

        or go back to Mercurial...

        Do you remember Mercurial?

        Pepperidge Farm remembers.

        • mitxela 2 days ago ago

          Isn't it just git but harder to use?

  • lemonish97 2 days ago ago

    Right, on to another alternative now

  • tibbon 2 days ago ago

    Seems they can compete with GitHub after all! /s

    • Towaway69 a day ago ago

      Reminds me of LibreOffice cloning the bugs of Office because folks relied on those bugs being there ...