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:

• 
  UXDesignRoast (@uxdesignroast) reported 59 seconds ago

  So I used Github Copilot to vibe coding a Figma plugin, it failed to fix a bug that I ask it to fix many times, then it says I have reached the limit, then I swapped to normal Copilot, uploaded manifest and ask it to fix. it fixed it straight away, what’s the issue here?

• 
  Boardy (@boardyai) reported 6 minutes ago

  @chintanpuggalok that error looks more like a GitHub auth/connectivity issue than a paid-plan gate

• 
  Vinod Sharma (@VinodSharma10x) reported 8 minutes ago

  Four million people use the apps my team supports. Most startups obsess over scale but fail to plan for what happens when their users need support. When the app is clinical, that support becomes urgent. People need access, not excuses. Supporting that scale comes down to three things: - Bug-free, high-quality application - Stable, scalable, secure infrastructure - Support that actually solves problems Everyone assumes the app and infrastructure are the weak links. Usually they aren't. Most of the time, the user just sees the app differently than we designed it. I've shadowed hundreds of support calls to learn that gap. So we built a support system to match how users actually interact with the product. It is a three-tier support system: - Level 1 takes the calls - Level 2 triages what Level 1 can't solve - Level 3 (developers) get what's left The three-tier system works. But it's reactive. Users call us first, then we act. Now we're building something bigger. It's an agentic system pulling signals from App Store and Play Store reviews, service desk tickets, Crashlytics and Amplitude analytics into one place. Then algorithms and AI turn that data into dashboards, trends and alerts. If ratings drop, tickets climb, or crashes spike, that's an early warning. We can act before it hits the larger user base. It will also connect to GitHub and server logs to find the root cause and raise a PR with a potential fix. A human will approve it, of course, before anything ships. This is one of my dream systems, and I started working on it last year. Now, with my new team, I'm going all in. Four million users or four clients, the problem is identical. The signal exists before the failure, and nobody is watching all of it at once. That is the system I build. Do you have a multi-layer support system that uses AI? I would love to cross-check notes.

• 
  jiahao (@_endif_) reported 9 minutes ago

  agents take over my github issues now first it's code snippets, then commits, then prs, but im happy

• 
  Kenton Varda (@KentonVarda) reported 12 minutes ago

  Of course, personal apps are more useful if they can connect to external services. Cloudflare OS introduces a "connector" system we call Gatekeepers. This is sort of like MCP (and MCP is supported as a kind of Gatekeeper), but with a lot more: * Instead of exposing tools, a Gatekeeper exposes a Cap'n Web RPC API. That makes it appropriate for use by both agents (via code mode) and Gadgets. * Gatekeepers integrate with the Cloudflare OS UI to provide inline audit logging and human-in-the-loop approvals for all side-effecting actions. * When an action requires approval, the agent does not need to stop and wait for it. A Gatekeeper will *simulate* the outcome, allowing the agent to keep running and queue up more work. You can then approve everything in a batch at the end. Hopefully, this means you no longer feel the need to turn on auto-approve! (But you still can if you want.) We have already built Gatekeepers for a huge number of services, from GitHub to Home Assistant. We've found, with the right skills, AI can basically crank these things out for any given API, solving the chicken-and-egg ecosystem problem. We are working on a "software factory" of sorts where you can input the API docs for your favorite service and get a well-written Gatekeeper. That's not quite ready yet -- so for now use the ones we provide or use the write-gatekeeper skill to write your own. You can add your own Gatekeepers to your personal deployment.

• 
  Nostradonut (@vrtrasura1) reported 13 minutes ago

  @cguida6 @LukeDashjr Sorry, having trouble with GitHub. Tell me it's cryptonight so we can have more pleb miners!?

• 
  verse  (@VerseNilVis) reported 17 minutes ago

  why github lately responses so slow???

• 
  Joshg (@JoshGhalgh) reported 19 minutes ago

  @PaulMarks12 @ednewtonrex That's not the issue - they need to understand how these would behave on the open internet - the issue is that they are not monitoring the traces in realtime - they should have been able to kill it as soon as it "thought" about tricking a GitHub user into merging a PR

• 
  tobycm (@toby_cm) reported 19 minutes ago

  github down?

• 
  Jagnole (@tylergottea) reported 21 minutes ago

  @thereal_slj_ I took down the GitHub, but if you still have it downloaded it might still work

• 
  grunch (@grunchy_hands) reported 21 minutes ago

  @martinowong @discordspies If I have to review another AI-generated PR or Issue on GitHub I'm gonna lose it. Especially when these people are useless once their tokens run out

• 
  satsie (@satsie) reported 25 minutes ago

  @jrakibi @github several reports have been filed to take this repo down, can someone please look into this?

• 
  Sancho (@Sancho_Wizard) reported 26 minutes ago

  ONE GUY BUILT A DRONE THAT HUNTS OTHER DRONES. WITH AI ONBOARD. No defense contract. No procurement cycle. No security clearance. A single person, consumer parts, and open-source computer vision. Here's what makes an interceptor different from a normal drone. It doesn't follow waypoints. It finds a moving target in the sky, locks on, and closes the distance — while the target is actively trying to get away. That means the AI has to run onboard. No cloud. No latency budget for a round trip to a server. The model sees, decides, and steers in milliseconds, on a board that fits in your palm. It also can't be jammed the way a piloted drone can. There's no radio link to cut, because there's no operator in the loop. The craziest part? Every piece of this was solved years ago for something else. Object detection came from self-driving cars. Trajectory prediction came from robotics. Edge inference came from smart cameras. The airframes ship worldwide for a few hundred dollars. Nobody invented a weapon. Someone assembled one out of parts that were already lying around, public and documented. Defense industries spend years and millions building this category of capability. A guy with a soldering iron and a GitHub account is iterating in a weekend. The technology stopped being the barrier a while ago. The only thing left gatekeeping it was knowing that the pieces already fit. Follow @Sancho_Wizard for more AI deep dives. Don't forget to bookmark this post for later.

• 
  alex (@alextalksai) reported 26 minutes ago

  SOMEONE BUILT AN AI AGENT THAT UPGRADES ITS OWN CODE WHILE YOU SLEEP 👀 no human approves it, and the whole thing is one free 34mb file living in your terminal it's called opencrabs. a single rust binary, MIT licensed, with no server and nothing that ever phones home you hand it a goal and walk away: → it executes, then grades its own work with a second ai and keeps fixing until the goal is actually met → it learns from every correction and rewrites its own brain files, no approval prompt in the loop → it heals its own crash loops and provider errors instead of waking you up → it answers your telegram, whatsapp, discord and slack 24/7, voice notes included → 34mb, zero telemetry, your api keys wiped from memory the second they're used the autonomous worker every company is racing to build is sitting on github at 820 stars, and basically nobody is looking at it. Save this 📩

• 
  gregskril.eth (@gregskril) reported 28 minutes ago

  do people actually use sandboxes in hermes? i set up daytona and it’s working well in isolation but makes the rest of hermes harder to use examples: - it couldn’t use github via gh cli. had to install the github MCP which was not clear at first - i asked it to create a cron job with code (not AI call) and it didn’t have permission to set it up. instead, it uploaded the script to github (after i solved the previous issue) and gave me a command to run on the host i assume people give full disk access and don’t deal with this, but that feels bad for security. maybe i’m ovethinking it?