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:

• 
  Alexander Zbiciak (@IBreakData) reported 53 seconds ago

  @luce_libera @wodarg One issue with the data is the data entry permutations are these the same lots? EW172 EW0172 PFIZER EW0172 #EW0172 Some entries just say pfizer or worse. The data needs guided cleaning to really show hot vs not lots @grok please review the github data and find the potential permutations for the pfizer lot ewo172 as shown above. Are there othe potential entries and what are the totals for that lot.

• 
  CyberDevHq (0xSEC) (@thecyberdevhq) reported 57 seconds ago

  @5mukx Dear brother, the best thing to do is to host your own repos via *** on your VPS is you have or want to, get the UI similar to GitHub and mirror it to the web Last year as the news dropped, that GitHub would now become a property of MS, I already knew many things will collapse. We are not ignorant of how MS treat issues reported to them, they either ghost you, or provide no serious solution for it. Look at their products, see how poorly managed they are. LinkedIn is an example I can never stop mentioning. People constantly complain, someone would open an account like today, the next day it gets suspended or restricted for no just reason. I said it a week ago I believe, let everyone host their own repos on their VPSs to avoid such unreasonable and unfair treatment. It's not gonna be easy but if this is how they randomly behave, you may wake up on day to find out that you've been banned. You can't ask why, because they do things without thinking twice, thrice. That's all I can say - prevention is better than cure. A wise wan foresees danger and dodge it

• 
  AIDegen (@nrsvv11) reported 3 minutes ago

  A 14 year old in China sold his first Python script for $40 on GitHub. The buyer turned out to be his own CS teacher. He did not find out until the first day back at school, when the teacher put it on the projector as an example of professional AI development. The kid was sitting in the third row. In America they are banning teachers from touching ChatGPT. In China a teacher just paid one of his own students for an AI agent and has no idea. He had built it over winter break instead of seeing his friends. Two weeks of asking Claude questions every night after his parents went to sleep. When the $40 came in he spent all of it on a Fortnite skin the same day and went back to coding. He pushed the project to GitHub with a README in broken English. ai agent that does homework and finds answers from any website. It sat at zero stars. He closed the laptop and went to dinner. GitHub Sponsors does not show the buyer's name. Just a username he had never seen. He did not care. The $40 was already a virtual outfit for a character he plays two hours a day. Then February. First class back. The teacher opened with a presentation on AI agents and ran a demo. A Python script that scans websites, pulls the data, summarizes it with Claude and sends structured reports on its own. I found this tool online and it changed how I prepare my lessons. It pulls from thirty sources in three seconds. This used to take me two hours every evening. The kid recognized everything. The variable names. The file structure. The comments he had left in Chinese because he was too lazy to translate them. His teacher was showing his code to forty students as an example of what a professional developer can build. He did not say anything. He went home and checked the fork count. 847. A university in Beijing had forked it to grade two hundred papers overnight. A tutor in Shanghai forked it into a homework checking service and charges parents fifteen dollars a month. A company in Hangzhou turned it into a support bot for an online store. All from a script a bored kid wrote over winter break with Claude. The forty dollars is a Fortnite skin. The code is running in three cities. His teacher still uses it every day and still has no idea who wrote it. The kid never told him. He said it would be too weird to tell your teacher that the tool he shows off to every class was written by the boy in the third row who still gets a B minus on the coding assignments. He gets the B minus for the code he writes in class, by hand. The A plus code is the one he writes at home with Claude. That is the one the teacher bought for forty dollars and presents as professional work. 847 forks. Three companies. One classroom that runs his code every day. He still sits in the third row. He still gets a B minus. Same kid. Same code. The grade just depends on who is looking.

• 
  Kleber Tiko (@klebertiko) reported 5 minutes ago

  @mattpocockuk I used yours with one good rule...every agent and context window must avoid hit 40%. If needed run /handoff with the summary and tracer-bullet tasks to next session. So i can run /clear and startover next tasks. My tracer bullet use github mcp server to create issues to track.

• 
  AIDegen (@nrsvv11) reported 7 minutes ago

  A Chinese developer posted a 41 second video on Bilibili showing the dashboard for his one person company. 19 autonomous AI agents handled everything 30 human employees used to do. Electricity bill: around 600 dollars a month. He pointed at the grid of cards on his monitor. Researcher. Copywriter. Designer. Developer. Sales. Support. Checker. The system ran 24 hours a day. He approved decisions from his phone on the subway. A laid off mid level manager at a Shenzhen electronics factory recognized the cubicle in the wider shot. He sent one screenshot to a Bilibili tech forum: timestamp 0:23. Pause at 0:23. Ignore the dashboard. Ignore the Dell monitors. Look at the giant union sign on the shelf in the upper right corner of the frame. That sign is not a decoration. That is the entry sign of his old employer's union office. The cubicle in the video was not his apartment. The cubicle was the corner desk in the union office where laid off workers came every morning for free coffee and wifi. He had not built a one person company that replaced 30 employees. He had been one of the 30. The company had laid him off six months earlier when they bought a SaaS platform that did eighty percent of what his department used to do. The 19 AI agents were real. The agents were also a demo. He had been running the system for himself for six months. He had pitched it to his former employer twice. Both times they had passed. The Dell monitors were not his. They belonged to the union steward's desk. He used them every morning from 9 AM to 1 PM because the union office had air conditioning and his apartment did not. The 600 dollars a month electricity bill was real. The electricity was the union office's. The union steward had agreed to let him plug in his local server in exchange for him helping ten other laid off workers polish their resumes. Six months ago a 14 year old in Shenzhen pushed an AI agent to GitHub. Judges said no real world application. 3,100 forks later. The laid off manager had been one of them. He had built the 19 agent system on top of that fork as proof that the company that fired him had been wrong about him. He was not a founder demonstrating the future of one person companies. He was the first laid off middle manager in his city to figure out the only way to win the AI replacement argument was to present yourself as the one who pressed the button. The clip is at 2.1 million views. The zoom on the union sign got another 1.6 million. Chinese tech viewers are still sharing the video. Still nodding. Still asking how to license the system. The system is still running. The cubicle is still at the union office. He still has not heard back from his former employer. He told the internet he had replaced 30 employees. The 30 employees he claimed to have replaced included him.

• 
  Julian Goldie SEO (@JulianGoldieSEO) reported 8 minutes ago

  𝗧𝗵𝗲𝗿𝗲'𝘀 𝗮 𝗳𝗿𝗲𝗲 𝘁𝗼𝗼𝗹 𝘁𝗵𝗮𝘁 𝗰𝘂𝘁𝘀 𝘆𝗼𝘂𝗿 𝗔𝗜 𝘁𝗼𝗸𝗲𝗻 𝘂𝘀𝗲 𝗯𝘆 𝟲𝟬 𝘁𝗼 𝟵𝟱 𝗽𝗲𝗿𝗰𝗲𝗻𝘁. Before your agent answers, it reads a mountain of text first. Every word is a token, and tokens cost you money. The more powerful the agent, the more it reads. This free tool is called Headroom. It squashes all that text down without losing the meaning. Same answers, but a fraction of the reading. In one test it crushed 10,000 words to 1,260 and found the same error. Nothing gets deleted. You can always unzip the full thing. You install it by pasting one GitHub link into your agent. Want the setup? DM me.

• 
  Atenov int. (@Atenov_D) reported 8 minutes ago

  GPT-5.5 + Claude Sonnet 4.6. Free $20 trial. No card required. Zed just hands it to you. Most AI IDEs run on Electron. Zed is built in Rust - rendering goes directly through Metal, DirectX, Vulkan. No Node runtime. No Chromium overhead. The result: responses that feel instant compared to Cursor or VS Code. >AI is built into the core of the editor, not added as an extension. That's the difference between native and bolted-on. Here's the setup: → zed(.)dev - sign in via GitHub (account needs 1+ month of activity) → download the IDE → click Start Trial → $20 in credits lands instantly, no card, no form GPT-5.5 and Claude Sonnet 4.6 available immediately. Zed also connects to Gemini, Ollama locally, and any provider via Open CDP. One more thing: real-time multiplayer collaboration built in natively. Like Google Docs but inside your IDE. No extensions, no plugins. $20 free. Fastest AI IDE available right now. One GitHub login.

• 
  WCNegentropy (@WCNegentropy) reported 9 minutes ago

  @JohnnyLitty2 @acadictive @theo That’s absolute bullshit and you know it. GitHub copilot’s been around since longer than A YEAR AGO and they’re JUST NOW running into these issues?? If memory serves me correctly, Theo himself made a video not too long ago about how GitHubs insane copilot demand caused outages.

• 
  Darth (@DarthSZNN) reported 9 minutes ago

  @ReaperDTX @Marskies_ DualShock GitHub, there’s all kinds of videos on it on YouTube. A circularity error of less than 7% is gonna make your sticks feel “stiff” in game and hard to micro adjust.

• 
  emeꕤ (@mixxariana) reported 11 minutes ago

  @uwukko hi wukko, are there any news related to drm? i was reading the issue in github and there hasnt been any new comment

• 
  HakanaiBlue (@HakanaiBlue) reported 13 minutes ago

  One of my friends hates that I complain about this stuff (our views just differ, still love him though) but I think things like this are A: Interesting and B: Good to report so programs can improve Also uh, is it weird that I sometimes just... browse GitHub issues for fun???

• 
  Micky Thompson (@mickythompson) reported 14 minutes ago

  @thdxr I use opencode from ~/GitHub folder and that contains multiple repos (packages, monorepos, and open-source projects). I do this because I often reference repos to emulate patterns in a project. Does that mean worktrees and more will not be available unless I switch to working from 1 repo? Is not working from a repo a bad practice in opencode?

• 
  Arpit Bhayani (@arpit_bhayani) reported 15 minutes ago

  SYN Flood is one of the oldest denial-of-service attacks, and it is still effective today. Here's what happens under the hood... A TCP connection is established with a three-way handshake: the client sends a SYN, the server responds with a SYN-ACK, and the client completes it with an ACK. What's interesting is that during this process, the server allocates memory for each half-open connection in a backlog queue. In a SYN Flood, an attacker sends thousands of SYN packets but never completes the handshake. The server keeps waiting for ACKs that never arrive, and the backlog queue fills up. Once it is full, legitimate users can not connect anymore. Thus, a DoS attack. What makes this attack effective is the 'asymmetry' - the attacker sends tiny packets with minimal effort, but the server has to allocate resources for each one. A single low-powered machine can overwhelm a much more powerful server. Fun fact: SYN floods have taken down GitHub, Cloudflare, and several databases in the past. To defend against SYN flooding, we can: 1. Cap the number of SYN packets from a single IP 2. Drop packets from known malicious sources 3. Or, the most effective, use SYN Cookies With SYN cookies, the server does not store anything. Instead, it encodes all the necessary connection information (client IP, port, and a timestamp) into the initial sequence number of the SYN-ACK packet it sends back. This sequence number is cryptographically generated, so it cannot be forged. SYN cookies make the handshake effectively stateless on the server side until it's fully verified, so the server does not reserve any resources until it knows the client is real. By the way, most modern operating systems have SYN cookie support built in. On Linux, we can enable it with `net.ipv4.tcp_syncookies = 1`. If you are interested, the Wikipedia pages are pretty well written for understanding this, and as always, you can use your favorite LLM to dig deeper.

• 
  m-ret (@mretsal) reported 17 minutes ago

  @GithubProjects We can't push to production on Friday because github is down.

• 
  蜃気楼 (@fuxps32) reported 21 minutes ago

  PewDiePie just shipped a free self-hosted AI workspace that runs entirely on your machine and your data never touches a server. No subscription. No tracking. No corporate model deciding what you can and cannot do with it. The thing runs any local model or connects to an API. Built-in agent that browses the web, runs files, handles real tasks. Deep research mode that reads sources and writes full reports. A memory system that learns how you work over time and gets smarter the longer you use it. The email assistant reads your inbox, flags what actually matters, and drafts replies in your own writing style. He described one automated reply as the most polite way to tell someone off they will never even notice. There is also a document editor, a calendar, an image editor, a comparison mode, and a tool called Cookbook that scans your hardware and tells you exactly what models your machine can run without breaking. All of it is open source and sitting on GitHub right now. The interesting part is not that it exists. It is what it signals. Every month the gap between paying for AI access and owning your own AI infrastructure gets smaller. The people who figure out how to build on top of free open source tools instead of renting access from the big labs are the ones who keep the margin. The tool is free. The knowledge of what to build with it is not.