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:

• 
  Jon Vorisek (@JonVorisek) reported 5 minutes ago

  @lydiahallie Hi Lydia, maybe your team could fix the issue where Claude Desktop for Windows refuses to relaunch after an update, forcing the user to reboot their entire workstation? It's been reported many times on Github without any response. Currently waiting to use Fable 5.1 due to that.

• 
  Pratik Patel (@PratikPatel_227) reported 6 minutes ago

  @github is down ?

• 
  Slim | ai-costguard (@SlimAssiliX) reported 6 minutes ago

  Three Anthropic billing changes this year. Three reversals before they landed. January. The OAuth block. Third-party agent tools stop working against subscription credentials. CI breaks overnight. Three days of forum threads and GitHub issues. Anthropic reverses it within days. June 15. The Agent SDK credit split. Programmatic usage moves to a separate per-token pool. Teams rebuild cost models. Anthropic pauses it on the day it was supposed to land. September 1. Today. The Sonnet 5 50% price hike. Teams audit prompts, evaluate alternative models, open migration tickets. Anthropic cancels it August 10. Three months of engineering reaction to three things that didn't happen. The failure mode is not Anthropic. The failure mode is that your production system has no runtime pricing awareness. Your agent made the same call through January's OAuth chaos, through June 15's credit split that never came, and through the September hike that was cancelled three weeks before today. It didn't know about any of them. It just called the API. Same request. Same response. No idea what the call cost relative to any ceiling. Every billing announcement reached your team through a blog post, a Slack link, a GitHub issue filed at 11 PM. The reversals came the same way. Engineers updated forecasts manually, from secondary sources, with an unknown lag from when the actual pricing page changed. Meanwhile, your agent was billing at whatever rate was live. Silently. Without surfacing the current price to any human on any call. Track cost per call in your runtime before the call, not in your calendar when the vendor sends an email. The GPT-5.6 Sol promotional window ends November 21. Whether OpenAI holds it or cancels it, your agent won't tell you. Your Slack thread will.

• 
  Theo - t3.gg (@theo) reported 8 minutes ago

  Just shipped a new T3 Code build with a bunch of small polish (and Fable 5.1 support!) Some of the changes I'm excited about: - Load model list from github (so we don't have to do releases to make new models available) - "settle" thread logic moved server side - Massive reductions in server-side memory usage in long threads - Massive reductions in idle CPU usage - Improved payload sizes for thread catch-up on client - Improved performance and reduced crashes on Windows - Hardened worktree creation/setup flow and status - Fixes for "stop" with subagents in OpenCode - Better image rendering behaviors - Tons more I'm forgetting

• 
  Kun Chen (@kunchenguid) reported 9 minutes ago

  @petergyang yo @myfirstmate peter just told me his skills are all at user level. backpass currently only runs things at project level i want a proposal for making backpass support a user level run. put that into a github issue use fable for peter

• 
  emrys (@eternalcloud003) reported 11 minutes ago

  @smellchone @transMossy stop i was having issues using github today😔

• 
  Jared Kubin (@JaredKubin) reported 13 minutes ago

  @ShanuMathew93 People are vibing everything. Honestly you have hackers, nation states, EVERYONE scanning GitHub right now for leaked api keys and unintended repo permissions This is THE problem with vibe coding … 95% is not being checked and major issues

• 
  Nav Toor (@heynavtoor) reported 13 minutes ago

  Drip is open source. Anyone can inspect the code. Its Google Play page reads: "Unlike other menstrual cycle tracking apps, drip is open-source and leaves your data on your phone, meaning you are in control." The code is public on GitHub. There is no company server holding your cycle data.

• 
  vshal (@snvshal) reported 14 minutes ago

  @shydev69 you can just spin up an auto pr-creating cron agent server if you wanna keep that github streak alive

• 
  Bhagya Mudgal (@BhagyaMudgal) reported 14 minutes ago

  @pcshipp can use github issues, agent work well with gh cli

• 
  Douglas Camata (@douglascamata) reported 23 minutes ago

  The "Notifications" page in @github is a great example of how not to do pagination and an UX totally broken by bugs and unexpected behaviors. First, the "inbox" doesn't have proper pagination not shows your all the grouped items. Marking notifications as read in a group is required to make the other groups appear. Then I click on one of the standard filters, "Participating". It paginates, showing me 3 items per page. There are 453 pages. I advance a few pages and suddenly there are 4 items per page. Few pages later, there are 5 items per page now. I chose the "review requested" filter now. It shows me 1 pull request per page. Page counter says "1-1 of 315". I move to the next page: it shows me the exact same pull request. I flip a few pages, it's still there. I go one page back and it breaks the UI completely, nothing shows up. What's going on?! See the video. There's no "go to page X", you can only navigate to the last (but not to the first). There's no page size configuration. There's no bulk operation being what you see in a single page.

• 
  Fluixo (@fluixoo) reported 28 minutes ago

  IF I HAD TO MAKE $13.4K WITH GROK BOT, I'D START HERE. Not with better prompts. Not by treating Grok like another chatbot. I'd give one agent one job people already pay for, one computer, one approval boundary and one measurable output. I mapped the full stack across 20 GitHub repositories: 1. the official xAI model, prompts, SDK and agent runtime 2. plugins and cross-agent delegation 3. open-source Grok Bot alternatives 4. remote control and multi-agent operations 5. MCP, search and infrastructure Grok Build gives the model hands. Grok Prompts exposes the behavior layer. The SDK and Cookbook turn it into a product. OpenMausBot and Rakazo give each agent a computer. HAPI and Herdr let you supervise the system instead of babysitting every message. The path to $10K is not finding a magic bot. It is choosing an expensive repeated problem, automating one measurable output, keeping human approval where mistakes cost money, and selling the result before adding more agents. The video captures the idea. The quoted article breaks down all 20 repositories, what each one does, the risks, and the order I would actually use them.

• 
  1mmモジュール (@1mm_module) reported 30 minutes ago

  @Eto_Madjas I’ve fixed the issue you reported. Please download the latest version from GitHub. I apologize that testing took some time, as the issue was related to the installation process. However, I don’t have access to an operating system with the same language settings as your PC, so I was only able to test the fix in a virtual environment. I hope this resolves the issue.

• 
  🌱Benson (@aduwaye77) reported 31 minutes ago

  @Krypto_Bishop @nvidia @100xengineers SkillSpector can scan skill folders, zip files, or GitHub URLs, running 71 vulnerability tests for issues like prompt injection and data theft

• 
  kkiran (@kkiran) reported 31 minutes ago

  @SparselyActive I had Claude look for a solution. I submitted an issue on GitHub for the author to take a look at this issue. I will continue monitoring this.