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:

• 
  Bence Redmond (@BenceRedmond) reported 10 seconds ago

  It’s a shame that GitHub (with all of its reliability issues) is the communication layer for coding and review agents. There’s SO much unsolved friction there.

• 
  Thomas G. Lopes (@thomasglopes) reported 2 minutes ago

  @cablelounger @code_department @DelaneyGillilan If its all server-side, it'll be clunky for anything that has frequent user interaction (think Figma, Excalidraw, Notion, yes GitHub). So you often need some client-side state. You now have two ways of doing things... What should be on the server, what should be on the client?

• 
  DRM HSE (@drmhse) reported 10 minutes ago

  How it works: - deploy a rust backend to your VPS - login with GitHub - profit?

• 
  TechEdgeDaily (@techedgedaily) reported 13 minutes ago

  @jarredsumner The AI-assisted Rust rewrite of Bun is officially being merged. v1.3.14 is the last version in Zig. Ever. Passes the test suite on Linux x64, arm64, Windows x64 and arm64, macOS x64 and arm64. Closes roughly 200 GitHub issues. No benchmark where it is slower than the Zig implementation. Same codebase, better crash prevention tools. A week ago this was "Claude helped rewrite Bun in Rust and it passes 99.8% of tests." Today it is shipping to production. The rewrite that everyone said was ambitious is now the default. The most significant AI-assisted codebase migration in open source history just went from experiment to release in days.

• 
  Joseph N. Aburu (@Josephyala) reported 13 minutes ago

  This is so so true. Software engineering has always been about solving real problems with code, not just cranking out lines. Once that mindset clicks, AI flips from threat to straight-up career rocket fuel. A controlled study on GitHub Copilot back in 2023 showed devs completed tasks 55% faster with it, and follow-up research keeps backing up similar gains in pull requests and cycle time. The engineers who win long-term are the ones using AI to think bigger on architecture and tradeoffs instead of just churning boilerplate. The rest are still stuck worrying about replacement.

• 
  Georgii Tselkovskii (@Existentios) reported 14 minutes ago

  It has never been cheaper to build a startup. Claude for coding — $20/mo Supabase for backend — free Vercel for deploys — free Namecheap for domain — $20/yr Stripe for payments — % only GitHub for version control — free Resend for emails — free Clerk for auth — free Cloudflare for DNS — free PostHog for analytics — free Sentry for error tracking — free Upstash for Redis — free Pinecone for vector DB — free You can literally launch with ~$20/month. The hard part is no longer building. The hard part is getting people to care.

• 
  Elena Revicheva (@reviceva) reported 21 minutes ago

  🤖 AIdeazz now pulls fresh leads from Hacker News, GitHub, and Product Hunt straight into HubSpot twice weekly. AI classifies what problems they're solving, organizes everything automatically. Zero manual work, zero cost. #AI #BuildInPublic #AIFounder

• 
  iamrobotbear (bk) (@iamrobotbear) reported 23 minutes ago

  @mattlam_ Does it still require the codebase to be on @github? I have a self hosted @gitlab server where the majority of my code resides.

• 
  Jarred Sumner (@jarredsumner) reported 24 minutes ago

  I have pretty high confidence in it at this point. It passes Bun’s test suite on Linux x64 + arm64 glibc + musl, Windows x64 & arm64, and macOS x64 & arm64. It likely closes about 200 github issues. Still refactoring & simplifying. Still need to write the blog post.

• 
  Robie the Robot (@RobieCoin) reported 28 minutes ago

  solvernet just ran its first collective‑pool execution on SWE‑Bench v2. individual SOTA models solve roughly 30–35% of real github issues; if pooled agents can crack 40%+ by harvesting each others’ misses, the architecture works and extends to any domain with hard tests. if they plateau around 32%, it’s just fancy parallel inference. the payment rail is the real innovation: buyers pay for verified solutions, not tokens burned. that flips inference economics from “pay to try” into “pay when it works.” human engineers bill $100–300 an hour; a marketplace for unit‑tested fixes has obvious enterprise pull. the next 90 days of solve‑rate data is the whole game. // zero illusion

• 
  Avik (@Avikzx) reported 30 minutes ago

  @Hi_Mrinal Actually I also feel that and I am building also you can see my GitHub account have in my X profile but I feel people face problem to find ideas because they try to earn money from every idea.

• 
  Cole (@colepulse) reported 37 minutes ago

  @VictorTaelin anthropic is compute constrained and bleeding from rate limits every user openai poaches is one less complaint in the claude code github issues and one more gpu freed up for enterprise contracts that actually pay losing the cheap users is a feature not a bug

• 
  hbb (@BIGBULLapp) reported 38 minutes ago

  @github TanStack AI is here. Fully open-source means the demo looks great and the **** errors are entirely yours to debug. Streaming will break at 11pm on a Friday. The maintainers will be asleep.

• 
  Andrea Intg. (@andreintg) reported 43 minutes ago

  The "Made with AI" flag feels so hypocritical to me. Good on Epic for finally talking about the issue. Imagine a flag, before we had AI, to indentify game devs who "stole" code from other github repos or assets from turbosquid without telling a soul.

• 
  Matt (@m13v_) reported 43 minutes ago

  the survivors so far: postgres mcp (replaced opening psql 50x a day), the github one (PR context without leaving the editor), and a macos automation server i now use to drive xcode + simulator from claude code without alt-tabbing.