GitHub Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

GitHub Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Li Bin (@OrNot2024) reported 1 minute ago

  @pyronaur @badlogicgames Same issue only for Codex but not for github

• 
  Josh (@joshmanders) reported 1 minute ago

  My AI assistant goal is to get my version in @dunnbot to a point that I'm managing it by creating github issues and reviewing PR's. When I approve and all checks are green it'll auto merge and deploy. The real kicker is that I'll be also using Dunnbot to plan and create the issues if they aren't bug reports from the platform support system, basically allowing me to go from coding to product management.

• 
  LV Nilesh (@LVNilesh) reported 2 minutes ago

  you don't really need a *** server. You can push code directly from one computer to another using SSH or local paths, without using GitHub or any internet-based hosting. *** remote add usb /path/to/usb/repo.*** *** push usb master

• 
  Craig Kerstiens (@craigkerstiens) reported 4 minutes ago

  @juliknl @mscccc One shared schema is generally quite fine, RLS gives you extra guaranteed enforcement. ORMs can be okay about protecting this, but defense in depth in terms of access to the data is not a bad thing. I know Mike we'll say we just had amazing developers at GitHub that knew how this worked and we didn't have problems. If everyone thinks they have a set of the best developers in the world great, but evidence is not everyone is that good and having something like the database protect you is not a bad thing. Things like actually separate databases per customer aren't inherently bad, but you need orchestration to manage across that. There isn't a right or wrong way. Sorry to not be more black and white controversial, RLS can be a good thing, a database per customer has some pros (and cons), but also 1 shared schema and customer data mixed together is totally fine and can work (but also has tradeoffs).

• 
  Penivera (@Penivera001) reported 5 minutes ago

  @github Please fix the branch export for codespace I have exceeded my subscription I was closed out before I could commit

• 
  SoGod | ☄️🌸(⚓🎀🐾)🎼👾 (@Gkay06_) reported 9 minutes ago

  @CsBuggy @cabanadrives5 Seems the GitHub pages have been taken down due to DMCA, so the one I've downloaded is the latest, to my knowledge

• 
  Awesome Agents (@awagents) reported 14 minutes ago

  A 755-point Hacker News thread and a 139-upvote GitHub issue document Claude Code Pro Max 5x users exhausting their quota in 1.5 hours. An independent investigation with 1,500 logged API calls reveals the math behind the drai… #ClaudeCode #Anthropic Link in the first comment 👇

• 
  basedcapital (@thebasedcapital) reported 16 minutes ago

  coffee, four terminals, one tab open on a vllm-studio github issue from three months ago with no reply. that's usually how you know it's still a solo project.

• 
  V (@madewithv) reported 18 minutes ago

  @github how about simply fix the uptime?

• 
  George Wizeman (@Moxorama) reported 20 minutes ago

  @github I would consider virtualised lists and a floating window - there's no point rendering 10,000 rows. The problem of editing a multi-megabyte file was solved back in the days of MS-DOS and 640 kilobytes, which was enough for everything.

• 
  Sublimation Chamber ⚗️ (@The_Sublimatory) reported 21 minutes ago

  @S1r1u5_ @gleech Yeah I think in a decade every time you you try to push, GitHub will just run a mythos+ level ai on it for free and if there are vulnerabilities it will give you a standard “failed to push” type warning and ask if you want to manually fix it or just let them automatically do it.

• 
  Vats (@VatsSShah) reported 24 minutes ago

  why MCP changes the shape of the problem: the agent doesn't just read "the user prompt." it reads tool results. every Jira comment, every GitHub issue body, every row from Postgres is input to the next LLM turn. you can't sanitise all of it. some of it is the content the user asked for. the only place to catch it is between the tool and the LLM- after the tool returns, before the content goes back into context. most gateways scan the prompt. most scan the response. the middle, where the agent actually gets exploited, is blind.

• 
  Andean Viber (@AndeanViber) reported 25 minutes ago

  @IntCyberDigest @0x686967 By default GitHub repos are Public... It's a major problem

• 
  Graham (@grah_uk) reported 25 minutes ago

  Real question as I scale my C#/.NET team: Do I:
– Add 2 juniors
– Add a senior
– Or double down on AI for the current team? We’re on VS + GitHub Copilot.
Cursor = insanely fast for prototyping
.NET = the bottleneck What’s actually working for you?

• 
  Incarnas (@_Incarnas_) reported 26 minutes ago

  Shipped an Obsidian plugin that syncs GitHub data into the vault. Smoke tested against real GitHub. Profile sync returning 404s on every repo. Paginated calls worked fine. Issues, PRs, releases, Dependabot alerts. All clean. Just the direct calls failing.