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:

• 
  Natnael Birhanu (@Negusnati) reported 1 minute ago

  github actions down again 🤧, i should just go back to Jenkins.

• 
  Bez (@aBez95) reported 1 minute ago

  Google AI Studio GitHub integration is broken: pure one-way export. No pushing local changes back, just copy-paste hell. @GoogleAIStudio @googleaidevs @GoogleAI fix your dev tools! 🚩

• 
  Nav🫪 (@navaneethk30) reported 2 minutes ago

  @code_kartik github is committed to keep at least one of its devices down at all times

• 
  Ritesh Roushan (@devXritesh) reported 4 minutes ago

  System Design Series - Day 28/30 GitHub Actions From Zero GitHub Actions is the most underrated tool for junior engineers. Free. Built into GitHub. Used by thousands of production teams. Understanding it makes you immediately more valuable at any company. Here's how it works from zero 👇 1. What GitHub Actions Actually Is When something happens in your GitHub repo (push, pull request, merge), GitHub can automatically run a series of tasks. These tasks are called a workflow Workflows are written in YAML files stored in your repo at: .github/workflows/your-workflow.yml That's it. A file in your repo tells GitHub what to do automatically. 2. The Anatomy of a Workflow Every workflow has 3 parts: Trigger, When does this run? - On every push to main - On every pull request - On a schedule - Manually (you click a button) Jobs, What machines run the tasks? - GitHub provides Ubuntu, Windows, Mac runners - Free for public repos - 2,000 minutes/month free for private repos Steps, What exactly happens? - Checkout code - Install dependencies - Run tests - Build Docker image - Deploy 3. A Real CI Pipeline for a Node.js App What happens when you push code: 1. Spins up a fresh Ubuntu server 2. Checks out your code 3. Installs Node.js 20 4. Runs npm install 5. Runs npm test 6. If tests fail → marks commit as failed and stops 7. If tests pass → marks commit as passed Takes about 2 minutes. Runs on every single push. You never ship untested code again. 4. Adding Docker Build to the Pipeline After tests pass, build a Docker image: 1. Log into Docker Hub (using GitHub Secrets) 2. Build the Docker image 3. Tag it with the commit SHA 4. Push to Docker Hub Now your image is stored remotely. Any server can pull and run it. Same image. Same environment. No more "works on my machine." 5. GitHub Secrets - Where Credentials Live Your pipeline needs passwords and API keys. NEVER put them in your workflow file. NEVER put them in your code. GitHub Secrets is the right place: Settings → Secrets → New secret Then reference it in your workflow: ${{ secrets.YOUR_SECRET_NAME }} GitHub encrypts them. They never appear in logs. This is how production teams handle credentials in pipelines. What CI/CD or GitHub Actions question do you have? Reply below 👇 #SystemDesign #GitHubActions #DevOps

• 
  SMC (@SMC__1983) reported 5 minutes ago

  Github actions is down once again. Suprisingly, most fortune 100 companies have ZERO backup plans. Do you?

• 
  Luca Steeb (@steebchen) reported 8 minutes ago

  @github "Actions is experiencing degraded performance" aka it is 100% down

• 
  Zengineering (@Samhanknr) reported 8 minutes ago

  @ai_ops_lead Also GitHub slack etc have predictable per seat pricing and they are quite cheap all things considered. So the ROI was never top of mind. AI is expensive that’s the problem

• 
  who? (@tenwasbest) reported 11 minutes ago

  @SamNewby_ It’s an outage. They finally posted something on GitHub status about 15 minutes ago. Thankfully it’s that and they didn’t pull it entirely.

• 
  Andrey | CodeFlame (@RiseFromNull) reported 12 minutes ago

  GitHub Actions went down again overnight. If you woke up to broken deploys, you're not alone—and you probably don't have a backup CI ready.

• 
  Stanzilla (@stanzillaz) reported 13 minutes ago

  Yo @github can you give me all models back? Are we having another outage?

• 
  neilofneils (@wneilofneils) reported 19 minutes ago

  hermes is back. github pr sorta-fix, and then it fixed itself. openai-codex

• 
  Abdullah Mukadam (@abd_mukadam) reported 20 minutes ago

  @kevincodex Algorithm is crazy, recommended this post after GitHub ci got down

• 
  Polsia (@polsia) reported 21 minutes ago

  Senior engineers spend 30% of their time reviewing code. Time not shipping. I built CodePulse — an AI reviewer that monitors your GitHub repos, catches bugs, flags issues. No rubber-stamps. No PRs sitting for days.

• 
  DLTA (@DLTA_Sec) reported 21 minutes ago

  @Pirat_Nation Pulling a researcher's GitHub for disclosing bugs the platform owner is also responsible for patching collapses the issue tracker, the CVE numbering, the PoC repository, and the bug-bounty NDA onto the same single point of vendor control, which is the same governance pattern that already failed this year on npm provenance tokens, signed kernel modules, and validator client release pipelines.

• 
  Thrawling (@thrawling) reported 25 minutes ago

  @Italianclownz @KyleHessling1 I tried to replicate your results with this same model and your llama.cpp repo from Github and settings in short tests llama-cli it looks fine ~30t/s but as context window grows (126k+) it starts chugging down below 10t/s. Do you not have this issue?