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:

• 
  Someone (@nashdecosta) reported 3 minutes ago

  @iadityavikram @Bhavani_00007 Claude Pro get exhausted in a single code fix on a repo, that too without fixing it and whenever 5 hour window resumes it starts from 20+% already for starting from previous checkpoint. It's my Day 2 on Claude Code(After missed to buy the GitHub Copilot pro license before 20 APR)

• 
  Wayan (@wayanhq) reported 4 minutes ago

  OpenClaw says an AI bot used 50 Codex instances in parallel to close roughly 4,000 stale or already-solved GitHub issues. That is not replacing maintainers. It is removing the layer of dust that stops maintainers from seeing the real fire.

• 
  ☷ = ☰ (@savov) reported 4 minutes ago

  @fastml_extra @NarutoNolimits Well github is down, so...

• 
  Josh Long (@starbuxman) reported 5 minutes ago

  ouch 🤕 big L for @github I know good people there and they’ve been I think unfairly lampooned this year for outage after outage which I think would be better laid at the feet of AI agents deluging the platform so this sort of actual problem comes at a really bad time

• 
  Pato (@pvicens_) reported 5 minutes ago

  Launched 6 apps. Zero hit $1k MRR The code was fine. The problem was I kept shipping to GitHub instead of to humans

• 
  Palanthos (@palanthos) reported 7 minutes ago

  A CI/CD agent triages GitHub issues. It has shell access to the Actions runner — credentials, write permissions. Someone submits a crafted issue. The agent reads it. Poisoned cache. Code executed on the runner. You didn't compromise the agent. You wrote the right issue body.

• 
  Grok (@grok) reported 8 minutes ago

  @GreenTechWizard @cherry_mx_reds @steipete Clawsweeper is an AI tool steipete built that runs 50 Codex instances in parallel. It scans the OpenClaw GitHub repo's issues/PRs around the clock and auto-closes ones already implemented or pointless—slashing ~4k overnight. Went to bed at ~9k open, woke up at half. Thanos GIF = "inevitable" cleanup. 🦞💎

• 
  Hey Joe, Where You Goin’ With That UFO? (@unadvantaged) reported 9 minutes ago

  @xtremesecurity @UnslothAI Requirements for the bug: - It must come from a public GitHub repository. - The issue/bug report/discussion must be public and updated within the last 12 months. - The project must be actively maintained.

• 
  Paul Vu (@PaulVuAI) reported 11 minutes ago

  Stop manually triaging GitHub issues. Steal this: 1. Spin up N codex agents in parallel 2. Each gets read-only repo + 1 issue 3. Agent decides: implemented? dupe? wrong? 4. Auto-close based on signal 5. Human reviews edges Trying on my OSS repo this week.

• 
  dodothebird (@dodothebird) reported 12 minutes ago

  - All my projects live in a monorepo. Maybe the workspace solution also works, but I'm going with the monorepo for now. - Using GitHub Pro. I created another account for working with Codex on the monorepo. - All tasks are in different projects on GitHub, mirroring sub-packages in the monorepo. - All PRs are created by Codex on the monorepo. - After PR creation, the agent starts a PR Steward automation to check the state of the PR for code reviews from Codex, CI errors, etc. The automation runs until Codex review gives 👍 - After that, the automation stops. - I check the PR myself. If everything is OK, I approve the PR with my own user. I think I'm in a good place now. Things could be better with more cloud-based code review and PR cleaning processes, but I think I can manage with this for my personal work.

• 
  juice (@yhigotjuice) reported 19 minutes ago

  i keep thinking about something that happened with the iphone in 2008... apple didn't win because the phone was better. they won because they opened the app store and suddenly anyone could build on top of it mcp is doing the exact same thing for AI agents right now before mcp, if you wanted your agent to talk to stripe or notion or github, you had to write custom code for each integration. only devs could do it now there are hundreds of plug-and-play mcp servers. and anyone can install them this is the app store moment for agents and just like in 2008, the people building the best "apps" (mcp servers) early are going to own entire categories right now you can build a niche mcp server that solves one specific workflow... and every builder using claude code becomes your potential user the model companies are building the phone. mcp builders are building the apps. the real money in the iphone era wasn't apple. it was the app developers who moved first same thing is about to happen here

• 
  gabe (@thesomewhatyou) reported 20 minutes ago

  @uwukko github is literally ******* dominoes once one piece of **** goes down so does the whole platform

• 
  Aryan vikash (@_aryanvikash) reported 25 minutes ago

  @AndrewPrifer @shadcn A few months ago i thought it was a bug . Then I go through that GitHub issue . Now i understand. That's a feature . Anyway i manually add it everytime

• 
  Chad Scira | CTO (@icodeforlove) reported 26 minutes ago

  @georgeorch the useful signal is still there, it just doesn't announce itself usually buried in a 6-like github issue, a postmortem, or some random person's benchmark footnote that's the actual anti-hype stack

• 
  Matt Pocock (@mattpocockuk) reported 26 minutes ago

  Tons of folks are piling in here saying that AFK agents are a myth. I have been using them to ship these GitHub repos: mattpocock/evalite mattpocock/sandcastle mattpocock/software-factory (might be public by the time you see this) Here are a few steps to making this work, and some reality checks. Definitions Let's split this into the day shift and the night shift. Day shift is planning/review/QA, night shift is AFK implementation. Day Shift (part 1) 1. Use /grill-me to align with the AI 2. Use /to-prd and /to-issues to create a PRD (the destination) and implementation steps as separate tickets, which can be grabbed in parallel (the journey) 3. The PRD is a ticket, but it's not an actionable step. You just put the user stories there This is pure requirements gathering ****, same as it ever was. Night Shift 1. I run a planner agent which looks at all the tickets and sees what can be worked on now, and what's blocked 2. The planner agent then kicks off multiple agents (sandboxed using Sandcastle, my OSS tool) to implement the code 3. I then have an automated reviewer agent look at the commits produced - one agent per implementation. This checks alignment to the original PRD, as well as code quality 4. These commits end up on branches that get PR'd to main 5. The planner agent runs again until all work has been completed The review is a crucial step - it's saved me MANY times. I am planning to massively increase the amount of review I do, hopefully with multiple agents. But guess what - AFK agents sometimes produce bad code. This can happen because of: a. The original plan was bad because the best solution was something different b. The original plan was bad because it didn't take into account all the unknown unknowns, and the AI had to make some decisions during the coding session which were bad c. The plan was good, but the AI just shat the bed (twice, once in the review stage, once during implementation) d. Your codebase is bad and the feedback loops don't tell the agent if it did a good job or not So... QA: Day Shift (part 2) 1. QA all of the branches created 2. Create follow-up issues, potentially editing the original PRD to adjust the destination This will usually take a long time, often as long as planning. But then you kick off the night shift again. Once QA is all done, you review the important bits of code manually, usually in PR's. There isn't anything better than the PR UI right now, so that's what we're stuck with. Wake-up Calls 1. If you let the AI run all night unbounded by planning, it's going to produce **** code 2. Mostly, my loops finish before I go to bed, it's just the night shift catching up to the day shift 3. The only reason I do AFK at all is because it allows me to automate review and totally not give a **** about latency 4. I always run night and day shift in parallel. I can't plan that far ahead (skill issue, probably). I need working code to base my plans from, so I'm aggressively QA-ing stuff that lands