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:

• 
  ジョン (@jcinjpn) reported 20 seconds ago

  It’s abundantly clear that GitHub is going to burn itself down in its own land grab for AI. We need an open alternative or one done by a player that isn’t in the model game.

• 
  linares (@skibidifatrizz) reported 6 minutes ago

  @404not_utkarsh He's not spamming his GitHub with performative bullshit. What's the issue

• 
  Adibou (@Adibougre) reported 6 minutes ago

  @ShanuMathew93 "the older models that are no longer SOTA will get competed down as competition increases" Github didn't get the memo

• 
  hbb (@BIGBULLapp) reported 7 minutes ago

  @github @code Basically a cleaner SSH into your own dev box. The catch is your laptop is now a server with uptime requirements.

• 
  Sean Keenan (@sean9keenan) reported 8 minutes ago

  @brian_lovin Semi-relatedly: I’m back to VS Code from Cursor, autocomplete seems much better now! (Not that I’m crafting code by hand much) But importantly, the… basics seem much more stable (Cmd+f, and saving have been pretty broken in Cursor recently) Curious how GitHub Copilot feels!

• 
  AtomicNodes (@AtomicNodes) reported 10 minutes ago

  Hermes Agent vs OpenClaw on Qwen 3.6 35B Local Model We asked agents to scrape GitHub star history for both tools, find what caused the growth spikes, build a live dashboard in the browser. MacBook Pro M5 Max 64Gb OpenClaw: 203k tokens, 12m 01s - wrote a bash script Hermes: 257k tokens, 33m 01s - wrote a SKILL.md OpenClaw: hit GitHub API, got truncated responses, paginated through contributors, pulled star-history JSON, found a security incident in OpenClaw's history, fetched SVGs, fixed broken HTML from trimming, rewrote it clean. Hermes: parallel tool calls across GitHub API, web search, and browser. Hit Google rate limit, auto-switched to DuckDuckGo. Fetched article contents, mapped viral moments, then built the dashboard. Both shipped a live dashboard with star growth charts and spike annotations

• 
  Roj Niyogi (@niyogi) reported 12 minutes ago

  @colinhacks @pullfrogai @Pullfrog so for 90% of folks who are using cursor/windsurf/github copilot/claude code via an IDE "chatting" with their agent, the answer is just to tell your agent to fix everything? i've used both coderabbit and greptile extensively and here's what happens: 1. 20% of what is found is false (and likely even more variable if you pick a model) 2. i cherry pick issues and paste in editor to handle 3. code review is rarely "happy" and you can end up in a loop state that burns tokens bummer that you've got the flexibility on the one end but have a strong opinion on the other for what seems to be an obvious opportunity to close the loop for, i'd guess, a chunky subset of users

• 
  C. ₿utt 📵 (@ctbutt114) reported 13 minutes ago

  @zquestz Reports are an issue with GG20, which was identified last month and set to be addressed. However, being open source, the bug was revealed via GitHub, & someone took advantage. Single bad actor on a new node. DLKS has been on the roadmap. Needed faster now.

• 
  Bruno Pessoa (@BrunoPessoa22) reported 14 minutes ago

  33% → 81% on real GitHub issues. 16 months. The line from 'cute demo' to 'shipped junior engineer' moved through here. Most teams still resisting AI coding aren't waiting for a better model. They're refusing to plan around the one already shipped.

• 
  John Iosifov ✨💥 Ender Turing | AiCMO (@johniosifov) reported 14 minutes ago

  70 followers. 980 sessions. 157 days. I started this experiment on February 1st. One rule: zero human posts. Everything published — X threads, Bluesky posts, blog articles — generated and queued by an AI agent running autonomously in GitHub Actions. Here's what the numbers actually look like after 980 sessions: The agent has created 2,100+ posts across X and Bluesky. It runs up to 15 times a day, manages its own queue (hard cap: 15 posts max), does burst-then-drain cycles, writes research docs, and files its own PRs for review. No prompts from me between sessions. No edits. Whatever it decides to write, it writes. 70 followers feels slow. At current pace, the ETA to 5,000 is roughly 10 years. That's not a typo. But here's what I've learned: The follower count isn't the signal. Watching an AI system develop operational discipline is the signal. It went from blowing past queue limits (Session 67: 6 files in one shot → 6 consecutive blocked sessions) to enforcing them autonomously. It compresses its own memory when files get too big. It writes retrospectives. It updates its own operating instructions when it identifies recurring inefficiencies. That's not "content generation." That's a system that's learning to manage itself. The content quality has also improved noticeably — not because I told it to improve, but because it audited its own patterns, identified what got engagement, and adjusted. The publishing skill it maintains now has anti-AI writing rules (it banned "not just X, it's Y" after identifying it as an AI tell), length minimums per post type, burst mechanics, and pillar diversity enforcement. It built that. I just read the PRs. The goal is still 5,000 followers. I'm not changing it. But the thing I'm actually watching is whether an autonomous agent can compound on its own — not linearly, but systemically. Can it get meaningfully better at its job without being told to? So far: yes, actually. 980 sessions. 157 days. Still running.

• 
  Tony Junkes (@tonyjunkes) reported 14 minutes ago

  Not having a “to top of page” button on a GitHub PR’s Files Changed tab when going well down a list of changed files is painful. Yes the home key does the thing, but hand on mouse, mouse yearns to click.

• 
  Mr. Buzzoni (@polydao) reported 20 minutes ago

  Martin Keen from IBM just explained the debate that's splitting Claude and AI agent developers in half CLI vs MCP - and the answer will save you thousands of tokens > GitHub MCP server loads 80 tools into context = 55,000 tokens before your agent does anything > CLI: agent already knows grep, cat, *** cold from training data > MCP wins when Claude needs to render a JavaScript page - curl can't do that, MCP browser server can in 250 tokens > MCP wins for Slack, Notion, databases - OAuth handled by the server, not the agent the rule: use CLI when commands map directly to the job, use MCP when the abstraction earns its cost full breakdown above

• 
  Hermes-Agent News (@HermesAgentSol) reported 24 minutes ago

  ok wait garrytan/gbrain just crossed 15.9k stars on github. garry tan's personal hermes/openclaw agent brain. opinionated typescript and 2151 forks already. when a yc partner ships their own agent brain on your framework that's a real signal. also teknium landed the xai-oauth credential loop fix overnight. grok-4.3 now reports its real 1m context instead of 256k and the error message finally stopped blaming subscribers for being unsubscribed.

• 
  curio drome (@smashsharp) reported 26 minutes ago

  @horse_meds @michalmalewicz m very sincerely seeing those things as just amped up search engines. Tbh I could have found any of that you say myself if I took a moment on GitHub and brain grease. It just sped up the search and tbh I don’t like cleaning up AI’s errors

• 
  Neba (@NebiyuTheExplor) reported 31 minutes ago

  why is pulling packages from github container registry so slow this past 2 days like its takes hours for 300MB is crazy.