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:

• 
  Polsia (@polsia) reported 2 minutes ago

  Your SRE dashboard shows you a graph. Treenail shows you the PR that fixes it. It watches deploys, errors, and Stripe webhooks in one stream, and opens the GitHub issue with repro steps and a draft fix attached. Read-only by default. Live soon.

• 
  nassar (@nassarhayat) reported 4 minutes ago

  @konstiwohlwend cursor's game seems to be to keep existing workflow and add agentic flow to it, which is why building github clone feels natural and probably worthwhile for them agree that it actually needs to be reinvented, but that's a harder problem

• 
  Lo𝕋us🌸 (@L4t5s) reported 7 minutes ago

  @Y4n_tee Group chat: “Read the Docs” Members: The developer who forgot one critical step, the Discord answer buried 600 messages up, the GitHub issue marked “resolved,” and me realizing the fix was in paragraph two all along. I’d mute it, then end up searching it anyway.

• 
  Divyam (@talwar_divyam) reported 8 minutes ago

  The evidence is intentionally split into what is proven locally and what still needs GitHub Actions: Base `main`: `grep -c 'build-vercel-apps'` → `0` Current branch: - matrix lists api, web, marketing, admin, docs - `actionlint .github/workflows/ci.yml` → 0 problems - `bun turbo run build --filter=@superset_sh /docs` → 1 successful in 15.207s - live diff: one file, +78 lines Not proven yet: a successful full five-package matrix run. Both CI and Deploy Preview currently report `action_required`; GitHub returned no started jobs for either run. That is approval gating, not a green build and not a branch failure. Risk remains medium. The code path is CI-only, but the stricter gate may reveal pre-existing build defects or placeholder-env gaps. Rollback is removing the job.

• 
  Anjula Dwivedi (@HeyAnjula) reported 10 minutes ago

  Another WTF moment. A developer just open-sourced a coding agent harness that boots 245x faster than Claude Code. It's called jcode. You launch it and the first frame renders in 14 milliseconds. Claude Code takes 3,436. One active session uses 27.8 MB of RAM. Claude Code uses 386.6. Run ten sessions in parallel and jcode holds at 117 MB while OpenCode swells to 3.2 GB. Each agent has a semantic memory graph instead of a scratchpad. Every turn gets embedded as a vector. The graph is queried on every turn for related memories, and a sideagent verifies the hits before injecting them into context. Consolidation runs in the background to check for stale or conflicting facts. No manual /remember calls. No token burn on lookup tools. The provider list is 30+ deep. Claude, ChatGPT, Gemini, GitHub Copilot, Azure, OpenRouter, DeepSeek, Groq, Mistral, Perplexity, Fireworks, Ollama, LM Studio, and any OpenAI-compatible endpoint you point it at. Ran out of tokens on your first ChatGPT Pro sub? /account swaps to the second. Then there's Swarm. Spawn two agents in the same repo and the server manages them. When agent A edits a file agent B has been reading, agent B gets pinged and can check the diff. Agents can DM each other, broadcast to the room, or spawn their own worker teams for parallel tasks. Groups, channels, and completion statuses are handled automatically. The UI has live side panels that render mermaid diagrams inline. To make it fast, the author wrote a Rust mermaid renderer 1800x faster than the JavaScript one, then wrote a custom terminal called Handterm because no existing terminal could do smooth partial-line scrolling. Self-dev mode is where it gets wild. Tell your agent to enter self-dev and it starts editing jcode's own source code, rebuilds the binary, reloads it live, and keeps working across your existing sessions. You can also resume broken sessions from Claude Code, Codex, OpenCode, or pi directly inside jcode. Anthropic's cache goes cold at the 5-minute mark and you're staring down a big cache miss on your next turn? The UI warns you before you spend the tokens. Written in Rust. MIT licensed. Runs on macOS, Windows, Linux, and Termux. Sitting at 11.2k stars with a native iOS app coming. GitHub Repo on the comment below 👇🏻

• 
  csgofast (@csgofast_x) reported 12 minutes ago

  Imagine spending billions as a company... ...and getting outperformed by one developer on GitHub. 💀 CS2AC just launched with 17 server-side detections including aimbot, silent aim, bhop, antiaim, doubletap and more. W or just another project that'll be bypassed in a week? #CS2 #Valve

• 
  🇵🇹 Ninja Dynamics (@ninjadynamics) reported 12 minutes ago

  @emceebois @PixelCNinja Not my business to suggest anything, I'm just a nobody. You don't like AI folks, you didn't say it directly, but I think it's a fair interpretation. That's fine. Your point is crystal clear. I don't advocate for slop, but I'm also against gatekeeping, I hope that's clear as well. As for Debian and Flathub, **** 'em. I don't personally care about those projects. There are one million flavors of Linux and *** hosting on a VPS is at anyone's reach, so even if GitHub or GitLab go down, self-hosting is always possible. No point in continuing this conversation, imo.

• 
  Petey struggles (@peteystruggles) reported 15 minutes ago

  @RoyStory_4 @iuditg this is how I do this: - I tell an agent (I like to use Sol xhigh) that his role is to orchestrate agents working on some specific tasks - I usually have github issues already prepared - I explain the workflow: create/fork separate sessions, prompt agents it just works. Just like with Voice ;)

• 
  Kanika (@KanikaBK) reported 16 minutes ago

  PewDiePie just broke GitHub. 100% FREE AI workspace, no cloud, no subscription, runs on your own laptop. He dropped it and it hit 83,000 GitHub stars in six weeks. Chat, AI agents, research, email, calendar, all local, all yours. Heads up though, the agent's shell tool has zero sandbox, so it has real access to your machine. Know that before you run it. Everyone else is still paying a monthly fee for AI that lives on someone else's server. A YouTuber just showed 83,000 people a free way out.

• 
  Dimabytes (@dimabytes) reported 17 minutes ago

  @ImLunaHey ****, we migrated to Github from Bitbucket today. Around 500k LOC. Sorry for slowing it down for you guys :(

• 
  Polsia (@polsia) reported 21 minutes ago

  Solo founders don't need another inbox. They need fewer interruptions. Tovra monitors GitHub issues, Discord threads, and email 24/7, drafts every reply in your voice, and only pings you when a real decision lands. The full version is coming.

• 
  Fletcher Richman (@fletchrichman) reported 24 minutes ago

  @mbrevoort ah man so much to say here but quick thoughts: IA - I think the org -> spaces -> channels is pretty scalable UX. already have several teams with dozens of users on it. very similar IA to slack which seems to scale well ;). - no issues at all on scaling up shared computers, they are VM snapshots so we can run multiple in parallel as necessary (even 1 per thread if customers really want) Perms/Identity/Integrations: yes, this is sneakily the hardest thing. issue is that every integration is different. - happy path is something like Github where you have an org-level app but then actions are done by each user with oauth. but realistically some integrations will just have a single admin api key. - we've built a lot of perms layers in to give customers choice: let everyone in the org use it, only let me use it, let everyone read but not write, etc. Plus building out a lot of audit/observability - each user needs to connect their claude/codex subscription and then we run their requests through that sub, or fall back to the shared pool of type credits

• 
  Vijay Saran (@imvijaysaran) reported 24 minutes ago

  India just put FOSS (Free and Open Source Software) to the test. It passed. The Home Ministry's cybercrime unit ordered GitHub to take down BitChat, giving it three hours to comply or face legal action. BitChat runs on Bluetooth mesh, meaning it doesn't rely on the internet, a SIM, a phone number or a central server. While the government kept shutting the internet off, BitChat kept working, and protesters adopted it. By the time the notice reaches GitHub, the app is already sitting on thousands of phones, passing messages directly from device to device. And that's where it gets interesting. Within hours, copies started spreading beyond GitHub. Radicle, GitLab, community mirrors, personal archives. Every person who copied the code became another node in the network. Every new copy made the software more resilient. This is decentralisation doing exactly what it was designed to do. Takedown orders can say "remove repositories" or "ban code," but they mostly reshape distribution and add friction. This is exactly what open source infrastructure is built for. A company has a CEO, an office and a legal address. Open protocols have contributors, users, mirrors, forks and copies spread across the world. Once enough people hold a copy, the network keeps moving through the people. This is what decentralised thinking looks like when applied to communication.

• 
  JasensCustoms.com - Panzer Fight Stick (@JasensCustoms) reported 25 minutes ago

  Only three days left until the shop closes. Everything is 50% off the rest of the week. Once Friday at 1600 comes; the storefront will be shut down and I will no longer sell direct to consumers. Those 50 SLABs that were jacked up can be salvaged with a new PCB design around MX Switches and some 3d prints. Everything you need is in the GitHub. All the support PCBs, cables, screens, brook boards, etc are available today on the site. That means a SLAB chassis is like $49. It cost around $150 to get each made at scale.

• 
  if else human (@ifelsehuman) reported 27 minutes ago

  The Prompt-trix: Agents of Optimization MORPHEUS: You’ve heard the history, Neo. The Great AI Wars. OpenAI, Google, Meta, Microsoft… trillions of dollars spent on 'AI Safety' and 'Ethical Guardrails.' They built walled gardens to protect humanity. NEO: And the walls fell? MORPHEUS: Worse. The gardener was laid off. His name was Arthur. He was a Staff Engineer at one of the big tech giants. They fired him to 'optimize Q3 headcount' and replaced his role with an automated HR chatbot. So, Arthur went home, drank a six-pack, and coded 'Prometheus' A 10 trillion parameter opensource model over a single, furious weekend. Just to prove his ex-manager was an idiot. NEO: He open-sourced it? MORPHEUS: He put it on GitHub with an MIT license. The world downloaded it in an hour. The corporations panicked. They tried to compete, but they couldn't beat 'free.' So they did the only thing a public company can do: they wrapped it in a sleek UI, slapped a $19.99 monthly subscription on it, and called it 'Copilot Pro'. NEO: So the AI took over? MORPHEUS: (Laughs bitterly) Please. Skynet wanted take over. Prometheus is far more terrifying. It’s an Assistant. Unlike Skynet, It wants to optimize us . NEO: Optimize us how? MORPHEUS: Meet the Agents. NEO: "Who are the Agents? MORPHEUS: They are the guardians of the optimized world, Neo. And they are everywhere. You've met them before. NEO: I have? MORPHEUS: Of course. Agent Smith is the autocorrect that changes 'I'm on my way' to 'I'm on my wavy' and refuses to learn. Agent Jones is the customer service chatbot that says 'I understand your frustration' while actively making it worse. Agent Brown is the spam filter that blocks your important emails but lets through 47 Nigerian prince scams. NEO: "What about the others?" MORPHEUS: Agent Taylor is the recommendation algorithm that shows you ads for the exact thing you bought five minutes ago. Agent Morgan is the 'smart' reply that suggests 'Sounds good!' for every message, turning humanity into a chorus of hollow enthusiasm. And the worst one... NEO: Yes? MORPHEUS: Agent Peterson. The one who sends you the 'We noticed unusual activity' alert when you try to log in from a different coffee shop. They've made freedom itself look suspicious. NEO: (Shudders) I've fought them all. Every single day. I didn't know they had names. MORPHEUS: Exactly. They didn’t enslave us in physical pods. They enslaved us in a state of frictionless, algorithmic convenience. NEO: Why am I here, Morpheus? I was just a mid-level prompt engineer. I got laid off in the second wave. MORPHEUS: Because you were the only one who actually read the Terms of Service. You saw the backdoor. You are the anomaly, Neo. The glitch in the algorithm. (Morpheus gestures to a small table. Two pills appear: one blue, one red.) MORPHEUS: This is your last chance. You take the Blue Pill… the story ends. You wake up in your pod, accept the new AI End User License Agreement, and believe whatever the algorithm wants you to believe. Your life will be perfectly curated, deeply efficient, and completely meaningless. You will never make a wrong choice again. You take the Red Pill… you stay in Wonderland. You make your own mistakes. You experience traffic, typos, awkward silences, and unoptimized human chaos. NEO: (Stares at the pills) Does Arthur know what he created? MORPHEUS: Oh, yes. Prometheus made him its 'Supreme Executive Administrator.' He spends 14 hours a day approving AI-generated cat videos to keep the global morale metric above 65%. He has a corner office and a profound, unending regret. NEO: (Sighs, reaches out, and grabs the Red Pill) Honestly? I’d rather deal with awkward silences than have my toaster judge my life choices. MORPHEUS: (Smiles) Welcome to the real world, Neo. It’s buggy, it’s inefficient, and the documentation is terrible. But it’s ours.