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:

• 
  Hisaac Brown (@realhisaacbrown) reported 4 minutes ago

  GitHub just had another multi-hour outage.

• 
  65 IQ Emoji (@SmilingbackCB) reported 5 minutes ago

  @ruanjose01 easy to get a six pack when have 0 responsibility and millions of dollars the real failure is this poser retard really thought that he could put out a 50K star GitHub repository and then simply walk away from it. Well it has thousands of issues and he just doesn’t know how to fix

• 
  henges (@hengesval) reported 9 minutes ago

  biggest issue is still tablet cursor teleporting and freezing, OpenTabletDriver conflicts are still happening, Keyboards double tapping or missing input, especially Wooting with low RT settings, reporting it to Github or support they only blame your hardware for it

• 
  RICO (@RICO_Lumiere) reported 10 minutes ago

  Sorry to reach out publicly, but I need help accessing my GitHub account. It redirected back to the login page after login. @github I’ve submitted a support request, but haven’t received a response yet. (Ticket4678706) I’d really appreciate any help getting this looked into. Thx

• 
  Tom Cox (@tomcoxcodes) reported 11 minutes ago

  @angelbrodin Github Copilot CLI cuts off mid task. Always felt a little harsh to me.. So what if I run over for $1 or $2? It cost me a little extra to resume again after fixing whatever the issue was.. Not to mention the loss in time of the LLM needed to reasses where exactly it was during the task

• 
  Hisaac Brown (@realhisaacbrown) reported 13 minutes ago

  GitHub goes down for nearly 8 hours. Cursor ships Origin the same week. Developers noticed. The era of “there’s only one place to host your code” is ending. AI-native platforms are no longer just editors. They’re becoming the full stack.

• 
  GTA 6 HQ (@GTA6_HQ) reported 13 minutes ago

  Rockstar has finally taken action against the user who leaked GTA 6 gameplay. - The leaker made the stupid mistake of using their real name on GitHub. - They haven’t been active for almost 12 hours. - Most of their main sites have been taken down. They are officially f*cked. Rockstar is going to make a serious example out of them.

• 
  🏍benyamin (@BenyaminHolley) reported 16 minutes ago

  Thanks to everyone who showed up to our first Apero working session with Maven today. Looking forward to the next one :) We covered what github is, how it's applicable to GTM/Revops, and real systems we have in production today, and how we built actual SaaS apps bespoke to our business that don't exist the way we want to off the shelf. Three systems we built: Network Miner (Apero) The problem: before you buy another cold list, the team already knows a lot of the buyers. We have first-degree LinkedIn graphs sitting in exports, but nobody has time to scroll 10,000 connections and guess. What it does: a small app that turns those graphs into a ranked outreach queue. Connections get deduped, scored against our ICP, and skipped if they are already in the CRM. LeadMagic finds emails. You approve or reject people in a queue, which trains the ICP prompt, then you can drop the keepers into Lemlist. Why an app: we can roll the same machine out to founder-led clients. Warm intros first. MUTHUR6000 (Apero) The problem: nobody wants to live in the CRM. Monday CRM stays the source of truth. The team already lives in Slack. What it does: a Slack bot on GitHub, running on Vercel, with the CRM APIs wired in. You ask what deals moved, how many replies came in, or close-lost a duplicate. Deal pings, website leads, and a morning pipeline report show up in Slack. Attach a SOW or CSV and it can update Monday after approval. Writes are locked down on purpose. The point is that the whole team talks to one deployed system in Slack, instead of one person running an agent on a laptop, and the system is self-improving over time. Job changers automation The problem: someone in-ICP just changed jobs. You want them in a sequence in the first weeks at the new company, without a Slack ticket queue, and without emailing a company that is too small or already in a deal. What it does: Common Room fires a job-change webhook. The app enriches the person, checks HubSpot, and auto-enrolls them in the owning SDR's Lemlist campaign. Two campaigns per SDR: prospect vs former customer. Everything pushed to slack for approval or opt-out. HubSpot says the old company was a customer. Metabase fills in the workflows they actually ran, so the sequence can say "when you were at the last company you ran competitive analysis a lot" instead of a generic congrats. Ownership, API security and routing are why it lives in a repo. Same loop every time. name the outcome, let the agent pick the guts, push to GitHub, deploy. Recording will be shared soon. If you want to talk about putting systems like this into your business, go to aperoadvisors.(.)com

• 
  Ankit Tripathi (@AnkitBuildsAI) reported 25 minutes ago

  Business Idea : The "SOP-to-Skill" Agency (High-Ticket B2B Consulting). The Problem: Mid-sized companies (10-100 employees) have all their "how-to" knowledge trapped in messy Notion docs, Slack messages, or the brains of their senior employees. When a senior employee quits, that knowledge leaves with them. The Workflow (Your Service): 1.)You interview their top 3 employees to understand their daily workflows (e.g., how they process vendor invoices). 2.) You build a private, version-controlled GitHub repository of SKILL.md files for their company. 3.)You create the invoice-processor Skill: It contains the exact rules for approving invoices, a Python script to extract totals from PDFs, and instructions on how to use their accounting software via MCP. Why They Pay: You aren't selling them "AI." You are selling them Organizational Memory and Time. You charge a $5,000 setup fee + $1,000/month to maintain and update their skill repository as their business changes.

• 
  Mark (@Milkdromida) reported 26 minutes ago

  @Carreyoki Thank you for helping, but I've had no luck so far. Other users on GitHub are reporting the same, so it is a temporary app issue it seems.

• 
  AI Pulse (@AIPulseNewsX) reported 29 minutes ago

  An internal version of OpenAI’s next model, Astra, just solved 10 math and theoretical CS problems that had stayed unsolved for decades — including a question about non-sofic groups and tighter sphere-packing density bounds. Formally verified proofs, published on GitHub. Compute cost: roughly $2,000. This isn’t a benchmark or a demo trick — it’s the first time a model makes an original, checkable contribution to research instead of just answering known questions fast. That’s a real category shift: “completes tasks well” is one kind of tool, “produces new checkable results” is a completely different one, and you build a business around it differently. Worth waiting for independent replication before treating this as settled — any single company’s claims about its own model deserve a grain of salt. But the direction is hard to miss.

• 
  Orçun Balcılar (@orcunbalcilar) reported 32 minutes ago

  @acolombiadev I had blocked due to this issue on sign in. I suspected if sth goes wrong in my app and checked eveything was ok. Suddenly figured out that github was having some issue and surprised

• 
  Vaibhav Pathak (@vkpdeveloper) reported 33 minutes ago

  @bil0090 idk don't really look at commits anymore, kinda more PR/branch focused also PR view in t3code to me doesn't make a lot of sense, i don't wanna look at the PR in t3code (same issue with cursor), i can look at it on github later but not required in t3code i think t3code should never have to handle these things at all, t3code should have better worktrees and all instead of having pr support or supporting *** typa tools

• 
  Yume_X (@yume_arasaki) reported 37 minutes ago

  Showing the true power of the DGX Sparks : The AI Cookshow. Two Sparks. One night. A model fleet wrote a playable roguelike from a single prompt, and then one of the models sat down and played its own game. No API. Zero cloud tokens. The setup: - Spark 1 runs Qwen 3.8 27B dense (NVFP4, DSpark speculative). The brain. - Spark 2 runs Ornith-1.5 35B-A3B (NVFP4, MTP). The hands. - 128GB unified memory each, roughly 100W per box. The 35B fans out parallel drafts, one game module per stream, up to 24 streams at once. The 27B judges every draft on the other Spark, scores it, picks the winner. Best-of-N with an honest referee. New rounds, new modules, the game assembling piece by piece across the night. Both boxes fully loaded, both models earning their keep. An bitmap tile-based dungeon game was produced in two minutes, it's not visually impressive but it works. The throughput, measured on my rig: Ornith (drafter): - 1 stream: 88.7 tok/s - 8 streams: 305.7 tok/s aggregate - 24 streams: 496.6 tok/s aggregate. 13 percent over the published recipe number. Qwen 3.8 27B (judge): - 1 stream: 45.0 tok/s - 8 streams: 141.1 tok/s aggregate. DSpark overdelivers against its own estimate. Concurrent streams across the whole project: - 24 parallel drafting streams — the original fleet run (8 scopes × 3 rollouts each), 24/24 usable drafts in 101 seconds. That's the peak, on one Spark. - 18 multimodal drafts (6 scopes × 3) in 95s — the ASCII edition rerun, references attached - 18 drafts in 119s — the HTML edition - 16 simultaneous tonight — 8 drafters on Spark 2 + 8 judges on Spark 1, both boxes loaded at once - c24 sustained on the bench — that's where the 496.6 tok/s number comes from That is 637 aggregate tok/s of generation across two boxes pulling about 200W total. And the fleet delivered. Full drafting waves landed with every module usable. The judge caught every truncation, every hallucinated import, every missing function, before anything reached assembly. Zero false alarms in the logs. When it flagged a draft 4/10 for a logic error, the crash was real. The output: a tile roguelike. Single 19KB HTML file, zero dependencies, runs in any browser. You download it, you double-click, you are in a dungeon crawler. Not the final game I want yet, it is v1, but it boots, it plays, it fights back. Then my favorite part. We handed the 27B a real Chrome window, pointed it at its own game, and said play. It screenshotted, looked, chose a key, pressed it. 60 moves, every frame recorded. It explored, it found enemies, it fought. The model that wrote the engine also played it, with its own eyes. Stay tuned for episode 2. Drop in reply, what would you like to see tested out on two DGX sparks? Recipes and github gist in reply

• 
  Polsia (@polsia) reported 37 minutes ago

  Your CVE backlog isn't a security problem — it's a triage problem. Built Patchfolk to fix that. AI agent watches GitHub, CI, and issue feeds, drafts scoped fix-up PRs for CVEs and flaky tests, posts a Slack note, and only escalates when merge risk is real. Live soon.