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:

• 
  Remy Cranen (@Remy_LeBeauBeau) reported 2 minutes ago

  @DannyLimanseta @alhuissi You can set up Actions on GitHub that basically run a check on the files in your repo, to scan for these kinds of potential security issues. You will get notified by the tool scan result. It's just another safety net. You can ask an AI agent to implement it for you.l on GitHub.

• 
  Somi AI (@somi_ai) reported 7 minutes ago

  Two compounding failure modes Datacurve's team surfaced. The tasks come from real GitHub issues, small and well-specified, built on data every coding model since 2024 has trained on. DeepSWE tasks edit 7 files and 668 lines on average. SWE-Bench Verified: 1 file, 10 lines. One is engineering. The other is recall.

• 
  Yashasvi (@_iamyt_) reported 7 minutes ago

  A sample DevMirror report: "Good morning 🌱 LeetCode: 3 problems · Streak: 2 days 🔥 Codeforces: Rating 1456 GitHub: 12 commits this week ✅ YouTube: 8 tutorials — build more, watch less 👀 📬 2 internship emails unread 📅 4/6 study sessions done"

• 
  Mediocre (@piyushpradhan07) reported 12 minutes ago

  Couldn't find anyone working on this in Github Issues. Thinking about vibe coding this over the weekend.

• 
  Tarak (@octavianslash) reported 12 minutes ago

  GitHub down again. The only way they can survive this Ai-pocalypse is to remove the free tier and to back to their old pricing

• 
  qualk 🌲🪓 (@qualk37) reported 14 minutes ago

  @threepointone @github made a 3CV and now your issues are gone

• 
  Vishal Pandey (@VP_code_n_lift) reported 15 minutes ago

  Question for Saas developers - Do you guys also create and solve issues in Github? - No ( We write the bug free code) - No ( We use notepad to write down any bug or enhancement observed) - Yes Me - I usually create issues in GitHub, if I or my user observe any bug or enhancement.

• 
  Greg Bergé (@gregberge_) reported 16 minutes ago

  GitHub is the shittiest software we are all obliged to use. When you have outage almost every day for a year you are just bad at engineering. Can’t wait for a replacement to pop.

• 
  CEO Govibe.org (@govibeorg) reported 17 minutes ago

  @piyush784066 I don't understand why sending your codes to a Microsoft server (Github) its like if you can not share any open source code outside of corporate control the open source community roasted me for not using Github saying that my codes were closed source cause its not on Github...

• 
  Natnael Birhanu (@Negusnati) reported 19 minutes ago

  @reckuza bruhhh why now when github is falling down do they want to move to it lol?

• 
  aigenerated (@abhxyyyy) reported 25 minutes ago

  github down again????

• 
  Soumava Das (@Soumava_221B) reported 25 minutes ago

  @github @GitHubCommunity Please look into this issue 🙏🙏

• 
  Hennie (@andyhennie) reported 33 minutes ago

  @theo Really good post!! We’ve all just accepted the shortcomings of worktrees (we should be able to check out the same branch in multiple places) and GitHub (monorepo with private folders please). There might be good reasons for things being the way they are. But the world has changed. We’re not working the same way anymore. The world must adapt. And people in charge can fix it. Whether it’s Linus or a GitHub VP doesn’t matter. We just need to point it out. And explain the need. I don’t always agree with you, and I sometimes get annoyed when I see you kick downwards on X. But at this ****, you’re really really good! Gj. If you get the issues on this list done (by Theo-bullying them or asking nicely, I don’t care), I’ll get a picture of you on the wall of my office. Probably one with a mustache. ;)

• 
  ed_the_engineer (@ed_the_engineer) reported 35 minutes ago

  Two years ago, open local LLMs were a hobby project. They solved 4% of real GitHub issues. The cloud was 20× ahead. Today a 9B model on a 16 GB consumer GPU solves 65%. The frontier sits at 80%. The gap that used to be a chasm is ten points. Look at the curve: → April 2024: Llama 3 70B at 4% on SWE-bench Verified → Jan 2025: DeepSeek R1 jumps the frontier to 42% — reasoning post-training arrives → Q1 2026: small-tier models (≤16 GB VRAM) cross 50%, then 65% in a single quarter → May 2026: frontier open ~80%, mid-tier ~73%, small-tier ~65% What used to require a datacenter now fits on a card you can buy at a consumer electronics store. What used to require the cloud now runs on your desk with the network unplugged. If the small-tier line keeps its slope for another 12 months, the model running locally on a $500 GPU will solve more real engineering tasks than the best cloud model could in early 2025. The question isn't whether local catches up. It's what we do once it has?

• 
  Sarah Yang (@sarahyang_00) reported 35 minutes ago

  @KettlebellDan This is a perfect example of how AI tooling is moving past the "chat window" era. For the last couple of years, developers had to constantly switch tabs, copy code blocks, fix formatting, and handle the GitHub pushes manually.