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:

• 
  Nate Matykiewicz (@NateMatykiewicz) reported 2 minutes ago

  @andreujuanc @mariorod1 @ThePrimeagen I just stumbled across it. Prime’s shared it (or at least screenshots of it) a few times. I’m a big GitHub fan, but the feeling of “something’s always broken” sucks. I know a lot of people assume GitHub’s use of LLMs is (at least partly) why uptime has tanked.

• 
  Gursahb Singh (@gs_rumana) reported 7 minutes ago

  And people are hating why GitHub is down. That's the reason bro. Every software has limits, GitHub has too.

• 
  Subhash Dasyam (@subhashdasyam) reported 9 minutes ago

  I found two security vulnerabilities in Anthropic's own CI/CD pipeline. Both are now fixed. Most breaches don't come through the front door. They come through the pipeline. The automated workflow nobody reviews anymore. The GitHub Action a developer added two years ago and nobody questioned since. The build step that quietly inherits every secret in your environment. CI/CD is the most dangerous trusted surface in modern engineering. - It runs on elevated permissions by design. - It pulls in untrusted code by design. - It has access to your cloud credentials, your API keys, your deployment tokens, all in one place. And - it runs automatically, without a human in the loop. That's not a bug. That's the feature. And it's exactly what attackers look for. The two issues I found in Anthropic's pipeline didn't require admin access. They didn't require insider knowledge. One triggered through a pull request. The other through an issue comment. Standard GitHub interactions that happen dozens of times a day on any active repo. Here's the part that should worry you. This wasn't just Anthropic's exposure. The same vulnerable Action was used by roughly 100 repositories across the open source ecosystem. Bun. Facebook's Hermes. Facebook's RocksDB. Rocket Chat. OrcaSlicer. And more. Every single one of them ran the same pipeline. Every single one of them inherited the same problem. One malicious pull request. Dozens of high-profile projects. An attacker could potentially have exfiltrated repository credentials or tampered with a build artifact. That's not a hypothetical. That's what a supply chain attack actually looks like in the wild. These issues were reported through Anthropic's bug bounty program and have been fixed. But the window existed. And most teams never know when their window is open. When did you last audit what your CI/CD pipeline can actually reach? What does it trust without checking? Who can trigger it and from where? That's the question worth losing sleep over.

• 
  Jeff (@AllenTheDetails) reported 14 minutes ago

  Using Claude Code (Sonnet 4.6) this morning to fix a simple bug in a process that runs on my GitHub actions. CC is on it's 5th iteration, has yet to fully resolve the issue, and I am now "approaching my usage limit". I'm on Pro. I pay for overage and am fine with that, but crossing into the additional pay realm after this amount of usage seems outrageous. @ClaudeDevs

• 
  Neo (@neohk852) reported 16 minutes ago

  @nmasc_ @ICONIQCapital Can you ask him to fix the billing issues for months that Anthropic is currently ignoring? people have been reporting about them for weeks on reddit and GitHub?

• 
  Osama Bin Chargin (@Amadeus3014) reported 21 minutes ago

  @TheHumanDurag I'll find a broken version on github. We can't let this opportunity float by

• 
  Super Human Epoch (@SuperHumanEpoch) reported 22 minutes ago

  @ashleymcnamara This is complete nonsense. Unadulterated victimhood mentality. Github isn't a scrappy startup trying to make ends meet and short of talent. GH is a multi-billion dollar business and mature product which companies and individuals of all sizes use and pay for. Y'all have massively screwed up not once, or twice or thrice. Last six months have been a comedy of horrors in terms of reliability. It's 100% on you to get your act together which is hurting the businesses that rely on the service and developers world wide. So please spare me the "we're human too" narrative. To err is human, software failures happen, take ownership of the issue and stop with the appeal to sensitivity.

• 
  dovvokun (@dovvokun) reported 23 minutes ago

  @matisanengineer GitHub is a product Dev's are using, and due to its constant problem it's causing an issue. Why can't they **** about the product that is not working optimally. Microvave was anology.

• 
  Roelf Diedericks (@RoelfDiedericks) reported 26 minutes ago

  @mitchellh I have also left github and shall never return. Host your own code. Its cheap. *** is a DVCS. Distributed Version Control System, by design. Github centralised distributed version control, and that's the entire problem. Self-host your own gitea, or forgejo instance, and federate.

• 
  Ejaj AHmed 🦅 (@aeejazkhan) reported 30 minutes ago

  🚨 Researchers Discover Critical GitHub CVE-2026-3854 RCE Flaw Exploitable via Single *** Push Critical GitHub Vulnerability Allowed RCE With One *** Push Short Points With Flow Security researchers discovered a critical GitHub flaw tracked as CVE-2026-3854. The bug affected both GitHub .com and GitHub Enterprise Server. It allowed an authenticated attacker to trigger remote code execution (RCE). Shockingly, the attack could be launched using just a single *** push command. This means anyone with push access to a repository could potentially exploit the issue. Researchers said the vulnerability was caused by a command injection weakness. If abused, attackers might run malicious commands on affected systems. GitHub quickly investigated the problem and rolled out a fix in under two hours. The company also said it found no evidence of active exploitation. The incident shows how even trusted developer platforms can face serious security risks.

• 
  abhay (@abhay) reported 31 minutes ago

  @jhiller @github Still? I thought that outage was over

• 
  Chidera (Di Maria) Humphrey (@ChideraCode) reported 31 minutes ago

  1. Find the stuck moment Go to GitHub issues, Reddit, Stack Overflow, and Discord. Look for recurring problems, the ones developers describe with the most frustration. Use keyword tools to expand the list. AI can help turn problems into cohesive topics.

• 
  🙎🏻‍♂️ (@ExxCP) reported 34 minutes ago

  @semihyesilyurt I am the app developer can you submit GitHub issue with the app log so i can fix it ? Or if you enabled shared storage disable it then try again

• 
  Mat (@matisanengineer) reported 37 minutes ago

  The loudest people shitting on GitHub are usually the ones who’ve never had to build or operate software at that kind of scale. Everything looks like a simple fix until you’re the one dealing with the complexity.

• 
  Ayush Mudgal (@ayushmudgal94) reported 38 minutes ago

  @punit_arani IMO, GitHub is in the midst of adjusting to the scale issues they’ve been having. Everyone just needs to give them some time to figure it out. I expect them to get back the four nines within the next 6 months.