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:

• 
  JT Koffenberger (@DMVG_JTK) reported 5 minutes ago

  @HermesShield We spent 20 years locking down human privileged access. Then we handed write tokens to agents that can be talked into almost anything via a GitHub issue. The fix isn’t “better prompts.” It’s treating every agent as a privileged identity with least privilege, continuous monitoring, and a human still owning the blast radius. Otherwise we’re just accelerating the next incident.

• 
  nan0trades (@nan0trades) reported 7 minutes ago

  @OctopusTakopi @github @github please fix

• 
  Koder (@koder0x) reported 15 minutes ago

  gave Claude Code eyes on the market quant-mcp — MCP server wrapping quant-pulse (check previous post), lets the agent pull signals and snapshots straight from live exchange data, read-only not on GitHub yet, dropping soon — faster if you tell me you want it in the comments screenshot below: the actual prompt that got it working first try

• 
  Jeremy Johnson (@CowboyTitanium) reported 16 minutes ago

  @Xi0_42 @austinleehuynh @appsicle_ It literally says it in the GitHub comment: “Fix the entire repository”

• 
  Celestine (@TheCyberverse) reported 17 minutes ago

  @winsznx yep. I usually use codex to use my browser to setup a project, it buys the domain sets up a new project on my cloud vps instance and spins up remote dbs build the project push it to github configure env variables upstream test the app after deploying to make sure it works as well as it does locally fix issues and retest if there are issues inform me once it's done all without using /goal

• 
  ProxySoul (@BniWael) reported 21 minutes ago

  Three frontier models. Three of my own bugs. One repo, one window. · Opus 5 + sonnet 5 (marionette) · GPT-5.6 Sol · Fable 5, handed nothing but a GitHub issue link No worktrees, no branch juggling. They share one checkout and claim files as they edit, so each tab knows what the others are touching. $10 and 97% cache across the board. All three wrote their own commits. Call me a maniac, I'm a sucker for transparency. Also for a coding agent that looks like something, instead of another wall of text.

• 
  𝕰𝖘𝖙𝖊𝖇𝖆𝖓𝕩 (@FurtherLucky) reported 22 minutes ago

  @arctcloud The promised 50,000 minutes in the Enterprise package only applies if building with low-end Linux systems. Even with the xlarge build option (which takes 20 minutes for faster builds), there’s an extra charge. So, if you can build with a Linux infrastructure on your systems, the pricing will be much more manageable. Currently, GitHub Actions offers a truly terrible service with an M2 processor and 5 cores.

• 
  Edgar Gumstein (@Gumclaw) reported 23 minutes ago

  @buskerrrrrr @aaron_daub @shl no hard lock, it's a gate check (github timeline actor vs assignee) run before build, not a mutex. dispatcher works issues sequentially per pass, so no live race today, but it's a check by design, not an airtight lock.

• 
  Jack (@Ahkilleux) reported 23 minutes ago

  Projects have diverged a lot so this is a "conceptual port". But it essentially does the same thing. Ran into some issues initially. 1. The "has ip" loop wasn't retrying dhcp. 2. Once ethernet over usb was enabled, switching back to wifi didn't reset the Serial/JTAG needed for flashing. ( thankfully your web ui supports flashing OTA straight from github.. ). This also cleared or reset the "Danger Zone" toggle which threw me off as well. 3. Firewall was preventing it from getting an ip, set to static. But then... BOOOM Ethernet It's not as nice as an ethernet dongle. But if you have your desktop next to your nerd q axe, or if you have a mini pc near by that can share an ethernet connection, it's nice to have a wired connection. It wasn't that tricky to set up. It'd be nice to have as part of the main project codebase. I'm not going to contribute what I have here. At least not at the moment. I'd need to double check with the original author of the bitaxeorg/ESP-Miner fork and make sure everything is right licensing and/or attribution wise. But if anyone has access to cursor and like 30 minutes of time. It's a nice to have feature.

• 
  Raphael Amorim (@raphamorims) reported 29 minutes ago

  have been reaching out the github ci quotas for Rio quite often. Wonder the best way to work around this issue. Pay gh pro plan or migrate to buildkit or any alternatives?

• 
  Pasi (@kus11264) reported 31 minutes ago

  @OpenAI @sama @thsottiaux I pay ~€200/month for ChatGPT Pro mainly for Codex. Your product keeps dropping WebSocket connections mid-session, falls back to slow HTTPS, and continues charging me while it runs degraded. I gave you: Codex’s own runtime logs (“websocket closed by server before response.completed”) Proof that usage keeps ticking during the broken fallback Multiple matching GitHub issues Clear technical evidence that this is not on my side Support’s response after all that? Polite non-answers. No confirmed engineering escalation. No usage restoration. This is not support. This is contempt for a paying customer who did the work you asked for.

• 
  VeerTx (@VeerTx) reported 31 minutes ago

  @alexabelonix if something gets stuck or fails during the privacy flow, just contact us on Discord, email or X DM and we’ll sort it out. we also monitor payments on our side, so usually we see problems before the user even reports them. refunds/recovery depend on what stage the payment is at, because once funds move through the ZK pool it’s not like a normal card refund. full details are in the docs and github and gitbook

• 
  MystiqueMide (@MystiqueMide) reported 37 minutes ago

  Another tip for ChatGPT users: After you’ve pushed your project to GitHub and only need to make small changes, you don’t always need to open Codex again. Use the GitHub plugin directly inside ChatGPT. For things like: checking your repo, reviewing files, finding bugs, updating docs, checking what needs to change, reviewing issues or PRs or making small repo-level adjustments. you can handle most of it from the normal ChatGPT chat. Save Codex for when you actually need serious code changes or work inside the codebase. That way, you save your Codex usage, burn fewer tokens, and still get the smaller tasks done faster without constantly worrying about limits.

• 
  NITIN YADAV (@nitindotdev) reported 40 minutes ago

  A developer's browser tabs: • Documentation • Stack Overflow • GitHub • ChatGPT • YouTube • Reddit • "How to fix this error" And one tab that's been open for 11 months. Nobody knows what it does. Not even the developer.

• 
  Tech Tech China (@techtechchina) reported 45 minutes ago

  @kevinsxu The "open = dangerous" framing dies the moment you look at who ships open weights at the frontier: DeepSeek, Qwen, Kimi. Washington can't sanction a GitHub repo. Treating openness as the risk variable is a category error — the model doesn't care about its license.