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:

• 
  Alex — Polymarket Odd (@oddsOnPMKT) reported 4 minutes ago

  Fable 5 Made Me a 5 Min Polymarket Trading Bot (full bot & results) High-frequency prediction markets look like easy money until you watch your P&L bleed in real time. I built a 5-minute strike sniper bot running live across four symbols—BTC, SOL, XRP, ETH—using lag arbitrage between Polymarket and external data feeds. The infrastructure works, the execution is brutal. Signals: 78% win rate on 9 trades (7 wins, 2 L's), daily P&L +$20.21, but two positions showing -5 each. One position down 73% with 16 seconds left, needed price under .72 in 9 seconds—missed by a heartbeat. ETH fill currently down 18%, another up 12%, hovering around 1828/1827 strike level. Position size: $5 per trade. Four simultaneous markets, 5-minute windows, taker orders for speed. Between the lines: the setup is solid—multi-account Polymarket integration, gamma API for market data, Hyperliquid for external pricing, automated share calculation based on size. But running four symbols at once in 5-minute markets is overkill. I leaned on Fable 5 to cook up the strategy from past GitHub patterns, which means potential overfitting to historical conditions that may not persist. No paper mode, live money on the line from day one. The more you trade, the less you have—frequency is the enemy here. Risk: sample size is laughably small. A 78% win rate means nothing after 9 trades. API latency, data sync delays, and competition in ultra-short windows can erase any edge instantly. Correlated losses across four assets would blow through capital faster than you can hit stop-loss. Daily stop loss is not optional here—it's mandatory. Flip: three consecutive L's, API latency spikes above 200ms, win rate dropping below 55% over 50 trades, increased order flow competition in 5-minute markets, correlated drawdown across multiple symbols, any change in Polymarket's fee structure or API limits. Trade: entry via lag arbitrage signals when strike price diverges from external data, $5 size until edge proven, 5-minute expiry window, invalidation if price does not cross threshold within 60 seconds of entry. Scale only after 50+ trades with consistent positive expectancy. Watch: API response times and data feed reliability, win rate normalization over larger sample, cross-asset correlation during volatility spikes, daily stop loss effectiveness, paper trading results before any size increase, market depth changes in 5-minute contracts. Bottom line: the infrastructure is ready, the edge is unproven. Start small, expect losses, and treat 5-minute markets as a laboratory, not a paycheck. #Polymarket #PredictionMarkets #HFT

• 
  Kosumi (@Kosumi1989) reported 7 minutes ago

  @efcon5 I always knew distribution would be the biggest challenge. Still, I built it because I needed it myself. I use it every day, so it's solving a real problem, not just sitting as another abandoned GitHub project. Even if no one else ever used it, I wouldn't regret building it.

• 
  Shredder (@Shred_0x) reported 8 minutes ago

  GET PAID UP TO $10 A DAY JUST FOR WAITING ON CLAUDE While AI thinks you earn money 1 Install the Kickbacks extension in VS Code 2 Sign in with Google or GitHub 3 Keep coding as usual When Claude generates a response a small ad appears instead of the spinner You get 50 to 70 percent of that ad revenue Heavy users report $1 to $5 on active days Cash out from $10 via Stripe You change nothing You just start getting paid for the wait Want the full step by step leave a comment under this post and I will send it to you

• 
  Kunal Dugar (@kunal_dugar) reported 9 minutes ago

  here we go again github is down

• 
  Momen Adel (@MomenBuilds) reported 10 minutes ago

  i was the kid who failed math finals and had to retake them in summer. the one teachers gave up on. in 7th grade, i was doing homework for 8th graders for 20 EGP each just so i could have money in my pocket. not because i was smart. i just understood what people wanted and delivered. i live with my mom and sister and see my dad a few times a year. not saying that for sympathy. that is just my life. i grew up knowing nobody was coming to save me, so i had to figure everything out myself. before i found building, i tried every internet money method possible. dropshipping, trading, random side hustles. i got scammed and failed more times than i can remember. but it made one thing clear. i did not want the normal path. then i found out one person with a laptop could build software and sell it. that changed everything. i touched code for the first time at 10 because i wanted to make games and could not afford to buy them. i learned from old youtube tutorials in broken english. at 13, i was vibecoding on a cracked windows laptop that sounded like a plane taking off. no github, no architecture, no idea how deployment worked. just ai tools and obsession. i started around 20 projects. most of them never launched. i would get excited, hit a wall, quit, then start something else. but every failed project taught me something about design, pricing, user flow and how products actually work. then a founder from the US trusted some random 13 year old from egypt he met on discord. he paid me $200, then $600, then around $1,700 total. that was the moment i realized this was not youtube motivation bullshit. it was real work, real trust and real money. i built keel ai, a tool that turns prompts into app mockups. i launched an app on the app store named after my mom. and at 15, i hit $4k in revenue from my dev agency because i refused to stop when adults smiled at me like i was some cute kid playing entrepreneur. some of the same people who told me to focus on school now dm me asking for advice. i do not reply to most of them. lately, i have been finding security holes in platforms people use every day. and it is always the same pattern. founders ship fast and leave their entire backend exposed. i have been on both sides of that screen. so now i am building something for vibecoders and solo devs who want to ship fast without leaking their users' data. not some boring security company with sales reps and enterprise calls. just a tool that catches the dumb mistakes before someone with bad intentions does. that is where i am now. still the F-grade kid who sold homework. still building on a laptop that might die any day. still refusing to stop even when it gets embarrassing.

• 
  wangfu91 (@wangfu91) reported 14 minutes ago

  @James_M_South @github I gave up on GitHub Copilot and cancelled my Copilot Pro+ subscription weeks ago after a series of incidents, bugs, and usability issues. Now, I am using Codex, and the experience has been much better.

• 
  Anushka Shandilya (@Anushka62255679) reported 14 minutes ago

  shipping the frontend and this is episode 9 of me building in public. what am i building? RAG for github that retrieves context not just from code files but also from PRs, issues, readmes and discussions. the frontend finally exists Last episode was all backend. This one I made the thing look like a product: two-pane layout, dark/light mode, a proper footer, and an octopus mascot called Inkling who follows whatever you're doing and points at it. I read Anthropic's "Introducing Contextual Retrieval" -Anthropic's core idea is that chunks often lose their meaning when they're embedded in isolation, so they first generate context for every chunk before creating embeddings. Reading that immediately reminded me of ReflexRAG because I was trying to solve a similar problem for code repositories by connecting code with PRs, commits, discussions, and version history instead of treating every chunk as standalone text. -one thing that really resonated with me was the shift in thinking. The paper isn't really about building a better retriever; it's about giving the retriever better representations of the data. It made me realize that retrieval quality depends just as much on what you embed as it does on which embedding model you choose. Anthropic's idea of contextualizing chunks before embedding also made me realize I'd been doing a lightweight version of the same thing in ReflexRAG. Instead of asking an LLM to generate context, my ingestion pipeline embeds repository-derived context like issue titles, file paths, source type, and line numbers alongside the chunk itself. The interesting question now isn't whether this works in theory, it's how close repository-derived context gets to LLM-generated contextualization. That's something I want to measure through retrieval evaluations. evals: my verdict on RAGAS Right now I evaluate by asking questions and reading the answers. That is not evaluation, that's vibes with extra steps. And it matters more than it sounds, because in episode 8 I changed: the fetcher (REST → GraphQL), the chunk budget (capped at 1500), the embedding checkpointing, and the citation pipeline. Any of those could have made answers worse and I would not know. I optimised performance for an entire episode and I cannot prove I didn't degrade quality. That's the honest state of it. i will discuss this in the next episode.

• 
  _iantyb (@_iantyb) reported 17 minutes ago

  @SingtelSupport Yes I did, GitHub is accessible after I first sent this message but now it is down again even after rebooting the ONT/ONR

• 
  Nathan Skene (@n_skene) reported 19 minutes ago

  UK is going to have to up it’s game in response to this: - Peer review is broken. Why are we not actively seeking replacements? - Grant applications are not as useful way of determining what big projects need to be done - Conferences / power point festivals should to be banned - Papers should be GitHub repos - Our institutions lack any systems for encouraging coordination / field building All the institutions of science need reform. America seems to have grasped this, what are we going to do?

• 
  N/A ʕ•̫͡•ʕ•̫͡•ʔ•̫͡•ʔ•̫͡•ʕ•̫͡•ʔ•̫͡•ʕ•̫͡•ʕ•̫͡•ʔ•̫͡•ʔ (@NA_outcast) reported 20 minutes ago

  day 225 -keep collecting data -fixed @github actions errors

• 
  Goshawk Trades (@GoshawkTrades) reported 23 minutes ago

  BitMEX is shutting down after 11 years. so it's worth going back to March 2017, a room in Hong Kong, maybe 50 people, and Arthur Hayes standing at the front with a python bot on the screen, teaching them how to market make bitcoin. the exchange had done one minute of downtime that whole prior year. it would go on to process $16B in a single day and over $1T annually. and here's the founder, giving away the starter code for free. but he tells the room the real thing: "you'll never see a very profitable trader handing out their market making bot for free on GitHub." and remember what BitMEX was at this point. it was one of the earliest and most influential builders of the perpetual swap, including the funding-rate mechanism that went on to become the standard across crypto derivatives. perps are now the most important instrument in all of crypto trading, and BitMEX helped define the exact structure the whole market runs on. the fact that Hayes broke this all down, this openly, this early, while helping build the thing that would shape the entire industry, is wild in hindsight. later on BitMEX did $16B in a single day and over $1T volume a year.

• 
  Meshak (@_Meshak) reported 23 minutes ago

  @petergyang Here I'll share few of my secret prompts : "I'm going to push the changes to Github and before that I want you to analyze all the current changesets in *** diff unstaged and ensure it has no bloatware, no boilerplate and no edgecases. Make sure the code is highly performant, production ready and easily maintainable. If not fix the code" "I want you keep this on your ****** memory, I dont want big sloppy AI code in my repo, I want Optimized code followed with best code principles to keep the codebase small , modular and clean. I dont want JUNK."

• 
  Dimitrios Prodromou (@dimitrios_pro) reported 24 minutes ago

  Following up on my mattpocock/skills post — here's what "using it daily" actually looks like when you're building a voice AI product: Every feature starts with /grill-with-docs. No code for the first 20 minutes. The agent interviews ME — edge cases, module boundaries, what happens when the caller hangs up mid-transfer. Annoying at first. Then you realize: every question it asks is a bug you didn't ship. The underrated part is the shared language it builds in CONTEXT.md. Our agents now say "containment rate" and "KB ingestion" instead of three sentences of description. Sounds cosmetic. It's not — the agent thinks in fewer tokens and names things consistently across the whole codebase. Misalignment is the #1 failure mode with coding agents. It was the #1 failure mode with human devs too. The fix is the same: talk before you build. Repo: mattpocock/skills on GitHub.

• 
  Jeff Huijsmans (@JRHuijsmans) reported 25 minutes ago

  Hey, if your *** clone with a PAT or GHP token doesn’t work, drop down to HTTP/1.1. That fixes @github ‘s vibe-breaking. For now.

• 
  Haydn Botts (@hxydnbuilds) reported 25 minutes ago

  Prompt: Act as an expert software engineer and UI developer. I need you to build a local, self-contained IDE application that visually and functionally mimics the 'Cursor' code editor, specifically integrating you (Grok) as the native AI assistant. This project is strictly for my own local device. Do not include any boilerplate for GitHub, open-source community guidelines, or remote repository management. Here are the strict technical requirements for this build: 1. Layout & UI (Cursor Clone): - Left Panel: A functional local file explorer tree. - Center Panel: A syntax-highlighting code editor with tabs for multiple open files. - Right Panel: A dedicated AI assistant chat interface natively utilizing my Supergrok capabilities. 2. Zero-Terminal Workflow: - The core feature of this IDE is that I must never have to manually open or type in a terminal. - Implement visual UI buttons (e.g., "Run", "Build", "Install Dependencies") that execute necessary background processes silently. - Surface any build logs or errors directly into the AI chat panel so you can immediately suggest and apply fixes without me touching a command line. 3. Local File System Integration: - The application must have complete read/write access to a designated local directory on my machine. - You must be able to read the context of my entire local project folder to accurately answer questions in the right panel. - When you generate code in the chat panel, include an "Apply to File" button that automatically writes your code changes directly to the correct local file in the center panel. Please provide the complete, step-by-step code and instructions to deploy this locally using a lightweight framework (such as Electron, Tauri, or a local web app). Focus entirely on making the local setup process as frictionless as possible.