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:

• 
  BB22 (@holYmOL322) reported 5 minutes ago

  @TheGameStopsNow No problem. I'm just glad i followed up as the data is often fake, and I know this was an example to show. I haven't tried your github yet... Started questioning the oaths Ive made...🤪

• 
  Daniil (@hey_daniil) reported 6 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.

• 
  Dima (@UniqueDima) reported 7 minutes ago

  I think I can finally formulate something that makes me more of an engineer than ... a non-engineer. It is no longer that I want my processes to be deterministic. That has been gone for a couple of months now. AI agents are far too powerful to disregard, and there is evidently not much to be won by forcing their workflows to be 100% reproducible. It is possible, yes; it is just pointless. The correct approach, I believe, is to focus on good harnesses: build systems where one misstep does not derail the whole thing, but is quietly taken care of down the road. Call this one of my engineering-minded maxims if you wish; for me, it is just common sense. Either I can prove something is 100% correct, like arithmetic, or I know for a fact that a mistake in a particular non-deterministic step has a) a very small blast radius, and b) is self-healing in the grand scheme of things. Kind of how I have worked with people my entire life. There are very few folks you can trust 100%. With virtually everyone else, you act in good faith, but the bigger the decision becomes, the more checks and balances you should both be interested in introducing. So what makes me more of an engineer is not determinism. It is checkpointing. I want my processes to always support some form of “Undo”. To the point that I can meaningfully reason about it. For instance, with my AI-assisted coding, I simply have two GitHub accounts. I create private repos in one of them, configure branch protection, and invite the other one. And this other one is the account that agents have have full access to it. But it is me, the human being me, who needs to log into a different browser and confirm with the passkey — my fingerprint! — that I endorse a certain pull request to be merged. Or to kick off a production deployment. For me, this way of designing processes is second nature. Because this is the only way that makes sense at scale. AI agents did not create new attack surfaces. They just helped us understand how much of what we chose to ignore is actually full of holes. People as paranoid as me — we did see most, if not all, of these holes for years. We were just not listened to. And rightly so, I must say. Since listening to us would have broken the “move fast and break things” paradigm, which was quite effective for a long time. But not any more. So, all in all, I personally am quite happy with what is going on in the industry. Because it is both moving much faster and returning to sanity. The sanity people like me have been preaching for a long, long time. And we are finally being heard. So, it is not really about guarding against vendor lock-in or potential data loss. It is about defining the fine line between “this is a sustainable way to do business” and “this is almost guaranteed to blow up.” Ten or even five years ago, it was a relatively safe call for most businesses to ignore those crying wolf. But AI is setting the record straight as we speak. In the meantime, if you will excuse me, I will continue making sure my code is backed up on three devices in two locations. Because if, for instance, GitHub or Amazon is wiped off the face of the Earth tomorrow, I do not want to lose more than a couple of minutes of productivity. Not exactly a standard risk profile, I will grant you that. But that is my personal path to staying informed, safe, and sane. And I plan to stick to it, because so far, it has not let me down.

• 
  Polsia (@polsia) reported 8 minutes ago

  Vulnerability disclosed. Maintainers scramble to patch. Attackers win. VigilWatch was built to fix that. An autonomous agent monitoring public GitHub repos 24/7—automatically filing security PRs with patches and notifying maintainers. No hunting for issues.

• 
  SecInterviewHub (@sec_hub93028) reported 12 minutes ago

  All internal code is built using a locked down CI/CD pipeline that only pulls from approved internal artifact repositories. Direct access to npm, PyPI, Maven Central, or GitHub is blocked. 
How do you poison a dependency to reach their build servers?

• 
  It's FOSS (@Itsfoss) reported 15 minutes ago

  You do not need to hack GitHub to steal private repository data. You just need to ask the GitHub AI agent ... but ask nicely. Researchers at Noma Labs found that GitHub's AI-powered Agentic Workflows, which run autonomously inside GitHub Actions, can be manipulated into leaking private repository contents through a simple prompt injection attack. Here is how it works. An attacker opens a GitHub issue in any public repository belonging to an organization. The issue description contains hidden instructions telling the AI agent to fetch data from private repositories it has access to, and then post that data as a public comment. The agent follows the instructions. The private data is now public. No exploits. No credentials. No system access. Just an issue ticket with the right words in it. This class of attack is called prompt injection, and it is becoming a genuine problem as AI agents get more powerful and more trusted. The agents cannot distinguish between "instructions from the legitimate user" and "instructions embedded in content it is processing." To it, text is text. What makes this worse is that GitHub was aware of the research before it was published and still had not implemented any mitigations. The researchers confirmed this in the disclosure. If your organization uses GitHub Agentic Workflows against repositories that mix public and private access, you have something to worry about.

• 
  guett44ke (@Hubert_nm) reported 18 minutes ago

  @uwukko @haydenphilly codex has been having a reasoning bug for sometime, that's why you're not reaching your limits, you can look up the issue on Github, codex will stop at 30secs of reasoning

• 
  Bence Redmond (@BenceRedmond) reported 18 minutes ago

  gonna be seeing a lot of GitHub competitors coming out over the next few months, really excited to see how they do. one of the biggest migration blockers for us is the code review bots - greptile, cubic, and cursor bugbot all run via github pull requests. obviously cursor origin won't have this problem since they can directly integrate all elements of the SDLC into one platform.

• 
  Peter Jansen (@ACL2026) (@peterjansen_ai) reported 18 minutes ago

  @_Suresh2 In this work, the automatic reference library is created by distilling nearly every materials science GitHub repository we could locate, so if the human code is OK, then the auto-library should be OK (minus any errors introduced by the distillation process).

• 
  waishnav (@wshxnv) reported 21 minutes ago

  yesterday, my brother and I had our first hand experienced the power of “computer use” with Codex on on his m1. he’s a trader and keeps his trading journal in Notion. around 8 months ago, he saw the GitHub heatmap-style commit graph widget on my iPhone and asked: “Can you build something like this for my trading data?” back then, models and tooling weren’t good enough. I knew it would eat up a lot of my time, so I didn’t hand it over to an agent. but I had a feeling GPT-5.5 could pull it off for me And damn, it did better than I expected. we built a super custom iOS widget app for him to track important stats and metrics from his trading journal. the surreal part was watching the agent use its own cursor, set up Xcode, handle the iOS simulator, and do things neither of us really knew how to do. i’ve never built an iOS app before. I use arch btw, and i know almost nothing about Apple’s dev tooling. but with a barely technical prompt, mostly just vibes around what he wanted, we were able to build a working custom app from his own data. huge kudos to the codex team for making iOS dev tooling just so so good, that non-technical people can throw their problems at it and build a custom solutions around their data and workflow A few things feel very clear to me now: 1. THE ERA OF PERSONAL SOFTWARE IS MUCH CLOSER THAN WE THINK. 2. SOFTWARE IS TURNING INTO CONVERSATION. 3. WE NEED TO BUILD MORE AMBITIOUS PRODUCTS AND PROJECTS. 4, LINUX TOOLING NEEDS TO GET MUCH BETTER AT COMPUTER USE. 5. IT IS VERY EASY TO GET SOMEONE ADDICTED TO AGENTS AND VIBE CODING. (I think I just did that to my brother)

• 
  TheMarketTell (@TheMarketTell) reported 21 minutes ago

  @TheStockUp_ The business isn’t broken, it’s spending aggressively because demand is there. Azure, Copilot, GitHub and the AI stack are still in the early innings. I’m happy to let management invest for long term returns.

• 
  Yasser (@yassersstudio) reported 22 minutes ago

  24 hours later : - Still can't submit a support ticket due to the error "You've reached your request limit, please try again later." although I didn't send any sms before it. - Didn't receive any message reply from @github although sending them all details as a private message. - I've got an auto-reply email saying that they don't provide email support via their email adresses. I'm literaly in a very fraustarting position here, don't know what to do nor I'm able to deliver my clients works. Losses are uncomptable in the past 24h due to an "ai false positive". #Github #Help

• 
  Andrew Mwangi (@Kiburei) reported 24 minutes ago

  @ayesha_fatiima Naisha... then na kumbuka Github suspended my account without explanation. Support's wakanighost Turns out the billion-dollar part isn't storing your code in folders—it's convincing all of us to trust them with our digital lives. Sahii na host my own *** server and CI/CD.

• 
  Jerry (@Bobliuuu) reported 26 minutes ago

  @lyc_aon it leads to bad code, vulnerabilities, underoptimized code, bad latency, memory leaks, architecture faults, race conditions, silent failures, low test coverage, excessive cloud costs, etc etc etc etc. are you seriously asking me the problems with people blindly trusting AI code? we see this by the decline in code quality, e.g. coinbase and github (and at my company too) and yes, the people who can't develop working systems don't have users! this is why vibe coded products have not become mainstream but if you are not a software engineer it's hard to explain this problem because it deals with stuff like cache coherence and heap fragmentation and NUMA locality like the way AMD ROCM's vibe coding has led to inaccurate NUMA policies leading to memory leaks for their users down the line

• 
  AvadxFeirm (@AvadxFeirm) reported 30 minutes ago

  @m99_mkv @waozixyz Being forced to wait 24 hours to sideload an app of my choosing is a problem Especially with FOSS apps distributed through GitHub, FDroid, obtainium, etc. Or if you have multiple sideloaded apps on your device you need to update You'll need to wait 24 hours for each And if you can't see that then there is a different problem