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:

• 
  Tech and AI Toolbelt (@TechAIToolbelt) reported 43 seconds ago

  @Koukyosyumei @ericlmtn Oh I figured it out. Had to use codex. Claude-mem got overloaded basically. Codex had moved some files around though in attempting to fix things. Got claude back open though enough to get a real fix and properly restoring my sessions, and getting backups to obsidian and github

• 
  Nico Gevers (@gevious) reported 2 minutes ago

  How are y'all working with AI code reviews? Has anybody got a workflow where you can do a code review in github/lab/etc and have the agent fix just those issues and push another commit for review?

• 
  Alexander (@richterich_) reported 3 minutes ago

  @github A good first issue is a gateway into open source

• 
  Joel 'Tukesi (@Eren_Jkc) reported 3 minutes ago

  A code push to GitHub triggers a Jenkins pipeline automatically through a webhook. Jenkins runs a lint check first so broken code never makes it past that point. If it passes, Vite builds the React app into static files.

• 
  Tubeguruji Youtube (@tubeguruji) reported 4 minutes ago

  @nigam13276 @browserbase Yes, will fix it asap, have you tried to connect github app first?

• 
  Bizzed (@BizzedAi) reported 6 minutes ago

  > Time to use those savings and an SBA loan to buy a liquor store and never login to GitHub again. but NOT before running your diligence on the store's business health and potential with Bizzed first 🥰🥰

• 
  Arka (@arkopra) reported 9 minutes ago

  @ydamitcodes For now I have not directly integrated X posting through URL params instead of using the X API as its paid. The LinkedIn one is automatic posting that you can use to Sign in through the settings. Give it a try! Even the GitHub auth only reads your commits and repo content. No actions taken otherwise.

• 
  @Saphyel:matrix.org (@Saphyel) reported 11 minutes ago

  @jakub_zalas Gitlab > github I mean github is always down lol

• 
  Amir Mansaray (@Increment_amir) reported 13 minutes ago

  @dkundel My browser use just doesn't work after it was renamed (open github issue) and sometimes /side does nothing. Codex half fixed the browser use bug, renaming from browser use to browser created issues. On windows

• 
  JH Trader (@JoshExile82) reported 13 minutes ago

  @1kartikkabadi1 @Teknium @NousResearch Okay, so its not just me. I did check their github and it does seem others are having the same issue too. So I assume it will be fixed or patched in time. Beside that, Hermes is a beast! Rarely ever have problems with it.

• 
  HumanOnlyWeb (@humanonlyweb) reported 14 minutes ago

  GitHub down?!

• 
  François Best (@fortysevenfx) reported 16 minutes ago

  @github Start👏by👏talking👏to👏the👏maintainers👏first. Opening a drive-by PR is likely not going to be the best experience for newcomers in OSS. If there's an issue: comment on it first. At least to say you're interested in working on it (locking it).

• 
  Babyface4Lyfe | VTuber (@Babayface4Lyfe) reported 16 minutes ago

  I’m seeing a lot of people be concerned about custom chat elements if stream elements goes down. I may look into developing a local file system that doesn’t rely on stream elements, that I’ll then put on github or something for the community. Currently, the stream elements thing is still just rumour and speculation so I won’t make any promises until I hear otherwise.

• 
  Overpower (@Overpowerfeed) reported 18 minutes ago

  @PrinceBuildsAI the "indie dev to openAI acquisition" pipeline is the most efficient data laundering operation in silicon valley history. they don't even have to scrape github anymore, you're literally just paying $20 a month to type your proprietary code directly into their training server.

• 
  Ben Anderson (@anchorstack_dev) reported 19 minutes ago

  Audited one of my old projects this week called PainChain. It comes from a time where I would let AI run rampant and generate features at will. Overall score: 2.6/10. Verdict: unsafe to operate. It is a change-aggregation platform for incident investigation teams. GitHub, GitLab, and Kubernetes connectors feed changes into a tenant-isolated timeline so teams can see what changed before an incident. The idea is good. The architecture is not nonsense either. The repo had JWT auth, OIDC, RBAC, session revocation, tenant guards, connector containers, Docker, docs, and a real attempt at a multi-tenant model. That is why the audit was more interesting. PainChain did not fail because it was a toy app. It failed because the production boundary did not hold. The biggest issue was simple and severe: Five of six API controllers were marked public. That meant unauthenticated callers could reach events, integrations, teams, projects, and timeline data. The integrations endpoint was the worst one. It returned connector config in plaintext, including GitHub PATs, GitLab tokens, and Kubernetes credentials. So the app had an auth system, but the valuable data lived behind routes that bypassed it. This is one of the most common failure patterns I see in AI-built and vibe-coded apps. The system has the shape of production software. There is a login screen. There is RBAC. There is tenancy. There are connectors. There is Docker. There are docs. But the actual boundary is broken and that is what matters. A login screen is not an auth boundary. The boundary is every route, query, token field, tenant check, and response body that handles real customer data. PainChain also had other production blockers: - production secrets committed to *** - API tokens stored unencrypted in the database - no database backups - schema changes applied with prisma db push at startup - no working CI for the v2 codebase - no tests on the current architecture - no health endpoint, uptime monitoring, metrics, alerts, or error tracking None of those are cosmetic issues. If real users connected real repos to this system, a basic network caller could read or modify sensitive data. If the database failed, there was no recovery path. If ingestion stopped, nobody would know. If a future change broke auth again, there was no test or CI gate to catch it. That is the gap between “the app works” and “the app is safe to operate.” A demo can work while credentials are exposed. A dashboard can load while tenant isolation is broken. A container can start while the database has no backup. A repo can look active while no automated check protects the production path. For PainChain, the first sprint should not be feature work, rather it should be containment: - Rotate exposed credentials - Remove tracked secret files from *** - Remove public access from data controllers - Encrypt third-party tokens at rest - Replace prisma db push with migrations - Add database backups before real customer data exists - Fix CI so every future change has a gate - Add a health endpoint and basic error tracking The product direction is still worth taking seriously. A change timeline for incident teams is a real product. The connector model fits the problem. The auth design shows intent. However, intent does not protect production data. Wiring does. The lesson from this audit: Do not audit whether the right features exist. Instead, audit whether the boundaries hold. That is the difference between a working MVP and software you can trust with real users, real repos, and real credentials.