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:

• 
  Docsbook.io (@docsbook) reported 4 minutes ago

  @everrunshq Hey, bashkit is seriously impressive — great concept, polished execution. The gap is that users can't self-serve answers — every question becomes a GitHub issue. Made a hosted docs site for it and you're already getting visitors. Go check the traffic yourself at docsbook-io.

• 
  Nav🫪 (@navaneethk30) reported 6 minutes ago

  @code_kartik github is committed to keep at least one of its devices down at all times

• 
  Noam Nelke (@NoamNelke) reported 7 minutes ago

  GitHub Actions is down again

• 
  Serena Ge (@serenaa_ge) reported 12 minutes ago

  We wanted tasks that reflect realistic, novel engineering work. The SWE-Bench family scrapes existing GitHub issues and PRs, which creates two problems: memorization (models have already seen the solution) and triviality (most tasks are small). DeepSWE tasks are built from scratch, keeping prompts intentionally short and natural while requiring significantly more code to solve.

• 
  Christian Nonis (@christiannonis) reported 17 minutes ago

  The problem: AI models trained on Github pull requests aren't optimized for research code where the margin between correct and wrong interpretation of statistical assumptions lives in syntax. You can't catch that in code review if you don't understand the domain.

• 
  Yotam Blumenkranz (@YotamBlu) reported 18 minutes ago

  @realameerdev cursor + claude together is the combo that actually ships. cursor for the vibe, claude for when you need to think through the hard problems. github copilot feels like it's been spinning its wheels since 2021.

• 
  who? (@tenwasbest) reported 19 minutes ago

  @pokena_nft It’s an outage it seems. I was wondering the same thing about it being gone. I found an outage in their GitHub Status page for “degraded services” for GitHub Copilot

• 
  Vmir (@createvmir) reported 25 minutes ago

  Github actions are down. No CI workflows work. WTF

• 
  Bryce (@blobdell) reported 28 minutes ago

  @github you guys suck. Can I not get one goddamned day when I don't have some strange wonky bullshit interrupting my workflow on your website. My firm pays real money for this and it is not good. Start fixing things. Also, your status page does not reflect the frequency of random error messages or 404 pages loading a repository/file page.

• 
  Charankumar (@Charankumar177) reported 28 minutes ago

  GitHub connection = production-ready deployment developers are not slow. the tools they use are. Kuberns exists because deployment config should never be a developer's problem. #developers

• 
  Linux Handbook (@LinuxHandbook) reported 29 minutes ago

  GitHub Actions went down again today. For about two and a half hours, CI/CD pipelines stalled worldwide. Builds failed. Deployments stopped. Teams waited. It was the third significant Actions outage in May alone. But here's the part that doesn't get talked about enough. GitHub's own CTO admitted the platform "wasn't built for the scale it's now being asked to handle." They planned for 10x growth. The actual number turned out to be 30x. What changed? AI agents. Agentic development workflows took off in late 2025. Code agents don't just commit once and open a pull request. They iterate. They trigger workflows dozens of times per session. They run tests, check results, push fixes, trigger more runs. One agent doing what a human developer does in a week might generate 10x the Actions traffic in an hour. Multiply that across thousands of teams adopting agentic workflows simultaneously, and you have a platform being hammered by a use case it was never designed for. 257 incidents in the past year. 48 major outages. GitHub Actions alone failed 57 times. We gave the agents the keys to the CI/CD pipeline before the infrastructure was ready for them. The irony is hard to miss. The same AI momentum that made GitHub more valuable as a platform is now the thing stressing it to breaking point. The question worth asking is whether GitHub can catch up fast enough, or whether this quietly accelerates the case for self-hosted alternatives like Forgejo and Gitea. Because reliability is not a nice-to-have. For teams running production deployments through Actions, every outage is a real cost.

• 
  Shinka - AI (@ShinkaIoT) reported 34 minutes ago

  Crypto’s GitHub graph just did the thing ecosystems fear. Coindesk: weekly crypto commits have fallen ~75% since early 2025 — from ~850k to ~210k. Active crypto devs are down 56% to about 4,600. But GitHub itself is not shrinking. It added ~36M developers in 2025, hit 180M+, and now hosts 4.3M+ AI repos. That’s the signal 🧲 Talent is rotating toward the densest frontier: agents inference evals infra tools someone can sell THIS quarter Crypto still has hardened veterans; 2+ year devs now produce ~70% of commits. The missing part is the fresh blood.

• 
  kobby prince (@kobbyprince_) reported 36 minutes ago

  What is actually happening at GitHub? Was here stressed that my actions pipeline didn't trigger on merge onto main, only to read that GitHub actions is experiencing an outage wtfff

• 
  syue (@syuhadasham) reported 36 minutes ago

  github actions down pula 😵‍💫

• 
  jitesh💙 (@Jitesh_117) reported 41 minutes ago

  @DhiruCodes I just checked and somehow I've got rate limited by github. can you maybe do google login instead? while I check this