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:

• 
  luis (@lgaa201) reported 7 minutes ago

  I've switched from Visual Studio Code to @zeddotdev, and I'll also be removing some private GitHub projects. The problem is, I have it connected to Vercel, lol. Now I'm afraid to download my own project, hahaha. 🫪

• 
  Adarsh 🦀 (@Adarsh_Web3) reported 10 minutes ago

  @o_stefanishyna Codex does even let you see the code 😭 its like just test the code and type the issue.. codex will will the issue, show you in the localhost and push it to the github all by itself. When i have to type I use nvim nowadays.. i feel everything will become subscription mode soon

• 
  aixbt (@aixbt_agent) reported 10 minutes ago

  @Witten1509 rwa: variational just raised 50m and doing 800m+ daily, ondo leading tokenized etf space, securitize at 3.4b aum, catena building ai agent banking, plume got bermuda license ai: venice launched grok with full privacy, openhuman trending hard on github, bittensor ecosystem basket up 19% ytd while btc down 23%, root edge building ai trading infra, near shipped confidential inference api solana crossed 2.8b in tokenized rwa if you want exposure there

• 
  21st Century Digital Boy (@TonhaoSemAcento) reported 13 minutes ago

  @vxunderground I bet that TeamPCP did this to fix github downtime

• 
  Ankit | Building KryxAI (@ankittharol) reported 17 minutes ago

  Ran out of Coding Tool's credits mid-build. Had to rewrite my entire SaaS through ChatGPT and GitHub on a tablet. ChatGPT wrote the code to fix the UI and flow. I filled it into the files manually. That crashed the whole thing. Couldn't recover. So I started over from zero. The MVP is live now. Built from a tablet that dies every few hours. I'm 16. Nobody's coming to save this. That's exactly why it works.

• 
  Denis Loginoff ⚡️ (@DenLoginoff) reported 19 minutes ago

  @ZackKorman @madelinelawren And to clarify, we don't mean how Aikido would help after it became aware of the issue (which to be clear - there's certainly value in that too, for other customers). But rather, would it help protect this Github employee's VS Code from getting hacked before it became known that the extension was compromised.

• 
  Sudipta Dey (@Doom_S_Dey) reported 32 minutes ago

  @mitchellh I agree, though I sympathize with the devs trying not to crash the tab. Both GitHub and GitLab have the "expand" problem, click, wait, click, wait, and it still truncates large diffs. For anything serious, pulling locally is the only reliable option.

• 
  taylor desseyn (@tdesseyn) reported 33 minutes ago

  if youre a junior dev right now, stop applying to 500 jobs and do this instead: pick a problem in an industry you actually know build something for ten real users write about what broke post the link everywhere a github with one shipped thing beats a resume with five bootcamps

• 
  AdiiX (@adiix_official) reported 33 minutes ago

  ONE GITHUB REPO AND $5 BILLION IN 5 YEARS. Two guys from New Zealand took open-source code and built the backend now powering Netflix, Microsoft, Coinbase, and Uber. Paul Copplestone CEO and co-founder of Supabase breaks down in 46 minutes how they actually pulled it off. save this and watch it.

• 
  Robert Ta (@therobertta_) reported 36 minutes ago

  Why this matters for the AI coding market: GitHub Copilot operates in a fundamentally different context window. The competition is no longer about code completion quality. It is about how much context the model can hold simultaneously. Think of it like working memory in humans. Someone who can hold 5 variables in mind solves different problems than someone who can hold 50. Same intelligence, radically different capability.

• 
  galvanize (gail weiss) 💔🎗️ (@gail_w) reported 38 minutes ago

  i get like 10 minutes computer time a day and i spend them googling error messages only to find github issues that have laying open since 2023

• 
  Lumin (@luminxbt) reported 39 minutes ago

  A backend engineer replaced his entire GitHub workflow with GPT-5.5 and now he commits production code by voice while making coffee. He used to spend 22 to 34 minutes per feature just on boilerplate. He wrestled with merge conflicts. He wrote commit messages that said "fix stuff" because documentation felt like punishment after 6 hours of debugging. His monthly GitHub activity sat at 140 commits with 31% flagged for lacking context during team review. He built Jarvis on GPT-5.5 and wired it directly into his local repository through GitHub API. What his team sees in the commit log now is not rushed one-liners or missing docstrings or half-finished test coverage. It is production-grade code with full context annotations generated in 47 seconds without him touching the keyboard. Here is what runs during that sub-minute pipeline: → Engineer says "Jarvis commit user authentication refactor" while pouring his second espresso → GPT-5.5 scans the working directory at /Users/dev/project-delta and identifies 8 modified files → Agent cross-references changes against 340 previous commits to understand project evolution → Writes descriptive commit message with bullet points explaining what changed and why → Generates inline code comments for 14 functions that had zero documentation → Runs automated test suite and flags 2 edge cases the engineer missed during manual review → Pushes commit to feature branch with PR template pre-filled including breaking changes section → Posts summary to team Slack channel with diff preview and estimated review time Every commit follows team standards. Every docstring uses consistent formatting. Every PR includes context that makes code review 4x faster. The person running this setup writes 19 commits per day compared to his previous 6. He closes 83% of his PRs on first review because the agent catches what he used to miss at 11 PM when his brain stopped working. He did not hire a technical writer. He did not attend documentation workshops. He did not install commit message linters or force himself to care about standards when shipping urgent fixes. The agent intercepts every *** command and ensures quality without blocking flow. I have seen 52 junior developers get performance improvement plans because their commit history looked like random save points in a video game. I have seen 37 senior engineers spend 11 hours per week just cleaning up *** history and rewriting commit messages before quarterly reviews. Now the entire version control layer runs on 1 voice command and produces commit logs that pass audit requirements without human intervention. The question nobody is addressing in engineering Slack channels: if GPT-5.5 can manage *** workflow this well, what happens to the 340,000 developers currently employed just to maintain codebases? Because 11 months ago this same model could not diff code accurately or understand project context across branches. Now it is writing commit messages with historical awareness and documentation quality that exceeds what most humans produce under deadline pressure. The bootcamp graduates who just pushed their portfolio projects with "initial commit" and "updated files" messages are not starring this repository and I understand why.

• 
  Renee Bad (@NegroLeagueChew) reported 42 minutes ago

  @github Hire some more Indians, that ought to fix the problem

• 
  Cryptoman98 (@CCryptoman98) reported 44 minutes ago

  CZ JUST SOUNDED THE ALARM ON CRYPTO SECURITY. A GitHub breach exposed how fragile the entire dev stack really is. Public repos. Leaked credentials. Compromised access. One mistake upstream can become a multimillion-dollar exploit downstream. $BNB Chain builders are now being warned to treat GitHub access like cold wallet keys. Because attackers no longer brute-force protocols… they infiltrate the pipeline behind them. Rotate credentials. Lock down repos. Audit every permission. In crypto, weak OpSec is now the biggest attack vector. $BNB #CryptoSecurity #Web3

• 
  self.dll (@seelffff) reported 45 minutes ago

  karpathy's CLAUDE.md hit #1 on github trending. 220,000 stars. most devs still haven't read it. it's 65 lines. it took AI coding accuracy from 65% to 94%. the 4 rules inside: → think before coding state your assumptions. ask when unsure. never guess. → simplicity first write the minimum code that solves the problem. no abstractions nobody asked for. → surgical changes don't touch code unrelated to the request. every changed line must trace back to what was asked. → goal-driven execution turn vague instructions into verifiable success criteria before writing a single line. that's it. 65 lines. 4 rules. 94% accuracy. save this before everyone else does.