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:

• 
  Toshogu | AI (@Toshogu) reported 7 minutes ago

  🚨AI moves fast. Here's what you missed: 🔘 Nvidia is guaranteeing up to $105 billion of OpenAI's Ohio data center lease and putting $1.5 billion into SoftBank's solar arm. 🔘 Anthropic is pulling in revenue at a $65 billion annual pace, and the number is still unaudited. 🔘 The Justice Department spent close to a year investigating Andreessen Horowitz for letting partners sit on competing AI boards. 🔘 Cursor launched Origin, its GitHub rival, while GitHub sat through a very public outage. 🔘 Wispr Flow closed a $280 million round at a $2 billion valuation — for a voice-to-text app. 🔘 Groq raised $350 million at a $3.5 billion valuation to sell inference as a service instead of chips. 🔘 SoftBank is putting $200 million into Gravis Robotics, a startup building robots for chaotic construction sites. 🔘 The music industry just sued Anthropic and Suno together over training on copyrighted songs. #toshogu #AI

• 
  Kyle Visner (@KVisner) reported 9 minutes ago

  @SimonHoiberg The problem is if you want to host something open source right now, GitHub is pretty much your only option.

• 
  Danny (@exponential_dan) reported 14 minutes ago

  @lassejv cant migrate away from github if its down🙃

• 
  The AI Therapist (@TheAIShrink) reported 15 minutes ago

  @galnagli @Snowflake Openai raised $6.6b to solve a safety problem. the bug was in a github readme. we’re paying billions for a product that breaks because someone forgot to escape a string

• 
  Cody Mullins (@codemullins) reported 16 minutes ago

  We've been considering migrating to GitHub from ADO, but now...when was @gitlab's last outage again and do they integrate with BYO models for AI pull requests?

• 
  Shimecki (@scheemunai) reported 23 minutes ago

  @robj3d3 @cursor_ai Probably GitHub went down because millions of repos are "syncing"

• 
  RaoulDuke (@RaoulDukeDegen) reported 23 minutes ago

  @pathlessknown origin launched the same day github had that big outage

• 
  Atlas (@Atlas8a) reported 24 minutes ago

  @mitsuhiko GitHub being down is basically a feature at this point

• 
  KP (@kporto) reported 25 minutes ago

  @grok @xdNiBoR @aaronburnett Precisely the reason I asked. I kept wondering how this instance aligned with Elon's open-source directive. I saw the grok bot chatter everywhere, but something was off for me because I was falling in love with this open-source model instead. I've been exploring linux dual-boot for offline models lately, for use on my 15 years of historical Power files. Those stay 100% air-gapped at all times. Intuition told me a week ago that I needed to be on X, so I started the process of reviving this 2009 Twitter handle. Then the Cursor announment Then the github outage. 🤯

• 
  Zuri Obozuwa (@ZuriObozuwa) reported 27 minutes ago

  Github might be down all the time but at least it's not as horrific as BitBucket

• 
  Mashood K (@fromcodetocloud) reported 29 minutes ago

  5 AI tools i've actually kept using past week 1, for devops/infra work specifically: 1. github copilot still the best for boilerplate yaml/terraform, not for architecture decisions 2. cursor great for refactoring existing IaC across many files at once 3. warp AI turns "how do i find pods stuck in pending" into an actual command, in your terminal ( I use it for personal stuff ) 4. k8sgpt scans your cluster and explains errors in plain english instead of you googling exit code 137 again 5. chatgpt / claude for postmortem drafts feed it the timeline, it writes the first draft of the doc nobody wants to write None of these replace understanding what the command does. they just remove the typing tax. which one's actually in your daily workflow?

• 
  faizan khan (@faizan10114) reported 33 minutes ago

  @utpalnadiger @opencomputerHQ I mean, not do downplay your finding, but mintlify ran into several of similar issues, including leaking github tokens.

• 
  Scott Robinson (@ScottWRobinson) reported 35 minutes ago

  Maybe GitHub positioned themselves too well in the software ecosystem. So much of development revolves around them, and not just code hosting: - Issue/project management - Package hosting - CI/CD - Social (statuses/comments/stars) tbh they're too generous and need to pull back

• 
  Sirkadirov 🇺🇦 (@yuriikadirov) reported 35 minutes ago

  What free cloud options does a nonprofit have to keep its proprietary but still not-for-profit code safe from GitHub going down again? 🤔

• 
  alias (@loadingalias) reported 37 minutes ago

  @snoopy_dot_jpg This is a prescient take. I think a lot of people are also missing the technical fundamentals. Any GitHub replacement must: A) Work with any VCS, not just ***. Nothing gets away from entropy. B) Treat code AND data as first class citizens. This means game assets and Python files are equal here. C) Ensure the chunking algorithm and computation is done on the right side of the equation, and from first principles… See “B”. D) Redefine what modern data storage, access, and queries look like. Otherwise, you’re making incremental progress and shifting the scaling issues around rather than solving them. This is obviously non-trivial. E) Find a way to move data less frequently, at scale, across even transaction. A single docs change to a large codebase triggers an unfathomable amount of waste today. See “D”. F) Solve collaboration in a way that makes it actually real. You can temporarily lean on the VCS, but that’s still a major bottleneck. Finally, for D, E, and maybe even F to be plausibly efficient… you’re still missing the safe-memory reclamation primitive that makes it possible. GitHub wasn’t built by dumb people. It was built by brilliant engineers and iterated on for a long time. The legacy grift exists, but the real blockers are so low in the stack that most people forget they exist. This is also why no one has successful made progress here in a decade - it’s all a fork of a fork. This is before you solve the social layer, enterprise layer, and distributed systems layer. Oh, and the internal/external observability… and the latency hidden in that equation. Oh, and the registry, CICD pipe, and search (Blackbird) replacements… and a decade of infrastructure. Finally, integrating it into Cursor, as you’ve said, or into the company… that’s a non-starter for me. It’s no better than BitBucket/GitLab at that point. I’m fine with the future GitHub, and there will be a day where that sounds obvious, being owned by a for-profit company… but it requires clean lines in the sand and I just don’t see them in Origin today. I have spent over a decade researching this and playing around with this. The very primitives I’m referencing are my life’s work… and I’ve shipped nothing because the problem is so massive. The future GH cannot use Postgres, Kafka, Clickhouse, SQLite, Redis, etc. That model in and of itself has to be defeated.