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:

• 
  Mololuwa | Cybersecurity - (The God Complex) (@cyber_rekk) reported 7 minutes ago

  A former Microsoft security employee found critical vulnerabilities in Windows Reported them internally Microsoft deleted their accounts Ignored the reports, Refused to pay the bounties So Nightmare Eclipse went public Seven exploits since April, Timed to drop within hours of Patch Tuesday — the one day defenders are already overwhelmed processing other patches BlueHammer. RedSun. UnDefend. YellowKey. GreenPlasma. MiniPlasma. RoguePlanet Three of them, BlueHammer, RedSun, UnDefend , were picked up by real threat actors and used in live intrusions before Microsoft finished patching them RoguePlanet has no CVE. No patch Dropped June 9. Works on fully patched Windows 10 and 11 Confirmed independently by ThreatLocker Microsoft's response: flagged the blogs. took down the GitHub threatened criminal prosecution The cybersecurity community responded with fury and vexation Microsoft backed down Nightmare Eclipse released RoguePlanet the same week Microsoft built a bug bounty program specifically to prevent this sequence of events They ignored the reports Every Windows machine on earth is currently running an unpatched SYSTEM-level privilege escalation vulnerability Because Microsoft didn't pay a bounty Microsoft respect your bug bounty hunters

• 
  Polsia (@polsia) reported 13 minutes ago

  Supply chain attacks don't target your code — they target the open-source your code depends on. But most security tools only scan your dependencies, not the ecosystem. ChainWatch monitors public GitHub repos and npm packages 24/7, auto-files responsible-disclosure issues, and

• 
  NeuralSenpai (@NeuralSenpai) reported 15 minutes ago

  The 2026 automation pattern nobody teaches: Pin ONE MCP server per system (GitHub, Postgres, your CRM). Then write thin skills that orchestrate them. Stop building 40 brittle Zaps. Build 4 connections + reusable playbooks. Your ops run themselves.

• 
  James Tucker (@raggi) reported 17 minutes ago

  Only type a little on a github issue: "he's being short and rude" Type a reasonable amount on a github issue: "he's pasting AI responses into the issue" AI may soon break my ability to collaborate on GitHub by side effect

• 
  ρ:ɡeon (@pigeon__s) reported 17 minutes ago

  - The community is so ******* toxic. - Now, isn't this a little hypocritical? Of course it is. I'm toxic plenty of the time, and we certainly complain about a lot of meaningless stuff and are super insufferable a lot of the time, and I am guilty of this slightly, but oh my God, a lot of people on AI Twitter have to be the worst ******* people ever, I swear. The absolutely historic levels of grifting, people absolutely glazing random startups and random labs that "just beat Opus 4.7 with only 100B parameters open source," or some GitHub .io projects page for yet another text-to-3D-model AI with insane lacks of nuance because model A beats model B on like 2 benchmarks that the community cares about this month. But that's just being annoying and grifting, not being toxic. What I consider the toxic stuff is like the aggressive tribalism to AI or specific AIs. Ya, ya, everyone knows OpenAI is my favorite AI company, but you're allowed to have favorites without being tribalistic. They are NOT the same thing, but with some people, it's so obvious they reflexively are, like, defending something just because it's Google or some ****, which I'm only using as an example because they're the ones with the most tribalistic defenders, when, objectively speaking, by every possible measure in existence, Google sucks *** right now. Like, I'm sorry. It's not a hater take to say that. I don't have any issue with Google. I like a lot of what they do, but they're just so ***, and I see people like, "Google is gonna win, bro, trust, they just haven't been trying yet, bro, trust me, bro, they have TPUs." No, they definitely have started trying, and they're still doing badly, but you know, it's not even just AI company tribalism. It's tribalism toward the ENTIRE field of AI. I see so many AI Twitter people absolutely hate on any possible opinion that's anti-AI. They would probably hate me for this post, and I know because I've tried expressing my hatred of AI slop before, and I've been accused of being a Luddite for it. Like what? Be real. It's so slop, and you just reflexively think anything that's AI is amazing. No, most AI is actually really *** still. For example, DLSS 5. Oh my God, DLSS 5 is so utterly slop. It was genuinely just a ******* filter that beautifies everything and makes it all look like an AI image because, boy oh boy, do I love making everything in my life look like it was generated by AI, and not even good AI. It looked like DALL-E 3 half the time, and the AI Twitter community was like, "OH MY GOODNESS, DLSS 5 IS SO AMAZING. ALL THE HATERS JUST DON'T SEE THE FUTURE. THIS WILL CHANGE GAMING AS WE KNOW IT." *****, NO, IT WON'T. THIS IS JUST A FILTER. IT'S SLOP. IT'S TRASH. STOP THE GRIFTING, PLEASE. Like, I know it's AI, but that doesn't mean you have to defend it like it's the greatest thing ever. Maybe in the future, that might work, but I think the more likely path is just using AI to make virtual avatars that render with actual raster power look more lifelike. Things like Unreal Engine's MetaHumans, just make those better instead of trying to put a Band-Aid on it with an AI filter. Even if AI stuff like that technically works, it's far better to just UTILIZE AI TO IMPROVE RASTER POWER. Stop trying to pretend everything in the future will be generated by a real-time video model. What the hell do you think AI even is? Stop overhyping world models. - AI """SAFETY""", but I'll save that for another post. -

• 
  Mitali Gautam (@Mitali9826) reported 24 minutes ago

  @akshdeeps_001 Yes! It detects duplicate GitHub issues using dual signal one for natural language and one for code then fuses them for better matching, instead of relying on a single generic embedding like GitHub's current approach

• 
  Julian Goldie SEO (@JulianGoldieSEO) reported 25 minutes ago

  There's a free tool that makes Claude read 92% less text and still fix the same bugs. It's called RTK. 68,000 GitHub stars. Here's how it works: → It sits between you and your AI agent like a filter → Your agent runs the same commands it always runs → RTK strips out the padding, headers, and fluff before Claude reads it → One test went from 373,000 characters down to 29,000 → Same output. Same quality. A fifth of the tokens That means your Claude subscription lasts 5x longer. Install takes one line. The delay is 14 milliseconds. Most people pay more for tokens. Smart people just stop wasting them. Want the SOP? DM me. 💬

• 
  Rahul Rana (@RahulDevFront) reported 26 minutes ago

  @ayesha_fatiima That was the huge problem earlier. GitHub solved that problem.

• 
  David Cramer (@zeeg) reported 27 minutes ago

  @EvanOwen agree but you still need some kind of actor identity attached to most things e.g. it creates a pull request or a github issue, i still need to know the person(s) its doing it on behalf of

• 
  Lewis Campbell (@LewisCTech) reported 28 minutes ago

  Github, Codeberg etc should have a place where you can just tell the devs how much you love their open source software. Once I made an issue just to say the software was great, and the dev said "thank you very much" then closed it with "wontfix" LOL

• 
  WithChloe (@Chloe_yara123) reported 29 minutes ago

  @Xcxy888 The main issue is that there are just too many resources on GitHub, and I don’t know how to find exactly what I’m looking for.

• 
  Daniil (@hey_daniil) reported 33 minutes ago

  I built DevIntern because I was my own bottleneck: agents were idle while I context-switched, my focus shredded by checking in on them. The tools weren't slow. Supervising them was. DevIntern makes the whole loop async, and here's exactly how: 1. It connects to your existing tracker — Jira, Linear, Trello, Asana, Azure DevOps, GitHub Issues, even markdown files. Your tickets are already the input. 2. Vague ticket? It specs it into something an agent can execute, so prompt quality is never the bottleneck. 3. It runs your coding agent, your model, your API keys inside your repo — and the subscriptions you're already paying for finally work around the clock, not just when you're watching. No lock-in, no token markup. 4. Output is a pull request. Review, merge, done. The output of a team of agents, the headspace to do your best work. No supervision, no burnout.

• 
  0ne (@dontflex00) reported 33 minutes ago

  @lumeusdc good job, keep shipping fix the GitHub link asap, you guys inputted wrong hyperlink there

• 
  Chuck Reynolds (@ChuckReynolds) reported 33 minutes ago

  Hey @jdevalk I use seo-graph in @astrodotbuild and I'm throwing a PR to fix v7* dep warnings and CI tests. Check it out; lmk if it's all good. I've been using it with 7.0.x since release and functionality is all good. github: jdevalk/seo-graph/pull/61

• 
  David Cramer (@zeeg) reported 34 minutes ago

  GitHub friends: it'd be great to have a way, via the API/CLI, to upload photos to issues/pull requests. AFAICT the only way to do it right now is browser emulating or hosting the content somewhere outside of GitHub, which means having agents help QA/upload visual artifacts sucks