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:

• 
  Cal Irvine (@Cal_Irvine) reported 3 minutes ago

  It’s my birthday so if GitHub wanted to go down for a few hours I’d probably be ok with it.

• 
  ©JΔΞOИ™✨ (@jasonshen_) reported 4 minutes ago

  Most voice agents recover by guessing. You cut in. It stops. Then it answers the sentence before yours. That’s not recovery. That’s a coin flip. StreamCore keeps the turn state. Duck -> Confirm -> Answer what you actually said. One Go server. Self-hosted. Apache 2.0. GitHub 👇

• 
  Kyrylll (@SojkoK) reported 4 minutes ago

  Trader with zero coding background built a full AI trading pipeline using Claude - scanner, entry filters, and backtesting Shay (Humbled Trader) connected Claude Code directly to her TradingView Desktop app and had it build her entire pre-market workflow in plain English. 🔘 The connection runs through an open-source MCP server (tradingview-mcp by @Tradesdontlie, ~9K GitHub stars) that lets Claude read live chart data and price action straight off TradingView 🔘 First build: a pre-market gap scanner - stocks up 5%+ from the prior close, priced above $3, over 50K in pre-market volume - auto-pulling tickers from Yahoo Finance and a one-line news catalyst from Benzinga for each 🔘 Second build: a strategy scanner that filters those gappers against her actual trend-continuation entry rules (price above yesterday's high, yesterday's close above the 200 SMA, breaking today's high) - turning a raw gapper list into an actual trade checklist 🔘 Then she had Claude write the strategy as Pine Script and back-test it inside TradingView itself: one ticker (MU) returned $1,200 P&L over 14 trades, a 9/14 win rate, and a 2.48 profit factor - before extending the same test across the "Magnificent Seven" Her caveat is the most useful part: "AI is mostly just slop without specialized knowledge behind it... you need real trading experience first, then AI becomes a tool to sharpen your edge" - not a shortcut to profitability. Note: the TradingView↔Claude connection currently requires TradingView Desktop on macOS specifically (she couldn't get it working on Windows), plus paid plans on both sides.

• 
  Tino Wening (@TinoWening) reported 10 minutes ago

  @Howaboua I’m pretty cautious when it comes to relying on third-party cloud systems. Think of GitHub, AWS, and CloudFlare outages. What happens if there’s an outage right when you’re in the zone? I think a local backup system makes sense if you don’t have much trust in third-party providers. It’s always a balance between dependence and self-maintenance. With the latter, you’re in control and can decide what happens, when, and how.

• 
  Doug (@DougStandley) reported 14 minutes ago

  At least I am not imagining this mess! Here is the response from @AnthropicAI: Fact, also confirmed — this matters most for your decision: This exact failure mode is a known, documented problem. There's an open GitHub feature request (#49649, filed by a Cowork power user in April 2026) describing precisely what you're experiencing: model upgrades effectively reset Cowork projects because specialist agents, accumulated context, and instruction-tuning don't transfer cleanly across model versions. The requester was staying pinned to an older Opus specifically to avoid this. As of that report, the capability to preserve a project across a model upgrade doesn't exist — it's a feature request, not shipped functionality.

• 
  Loftwah (@loftwah) reported 14 minutes ago

  Finally hit by GitHub issues.

• 
  Mudkip 🎬 (@MudkipOnYT) reported 16 minutes ago

  @ReticentY2K It’s cool but still very early days, time trials ghosts don’t sync so while RWFC is down it’s quite pointless sadly, it also doesn’t support Wiimmfi in the base game. Issues are being raised and looked at on GitHub so I’m hopeful Patchzy will get it sorted, lots of potential.

• 
  Simon Skinner (@vultuk) reported 17 minutes ago

  @Rames_Jusso Nope. I have a Mac mini that just runs codex. It’s just sat there dealing with things for me. Runs scheduled tasks to work on GitHub issues and I connect through remote to start it working on things. It’s more than enough. No usage issues, no need for anything more. There seems to be a whole batch of people that seem to “need” all these special agents, yet they aren’t shipping anything that benefits from it. (That I’ve seen at least)

• 
  Andrea Fomera (@afomera) reported 17 minutes ago

  Tried to checkout @laravel cloud today, some weird GitHub app issue blocked me. Met the AI response bot with an entirely unhelpful response. Nice.

• 
  Frooxius - frooxius.bsky.social (@Frooxius) reported 19 minutes ago

  @SweetHeartKnows Dude, that sounds more like you got an axe to grind. Sure there are issues, but you're being a bit extreme. We've been making a lot of imprints to these, that have been helping. If you got any suggestions, our GitHub is open.

• 
  Big Pip (@Model_Culture) reported 22 minutes ago

  AI CODING AGENTS ARE BREAKING ONE OF THE OLDEST SIGNALS IN SOFTWARE HIRING: THE PORTFOLIO. A girl in China built a project combining computer vision, hand tracking, interface work, backend logic and an AI agent. A few years ago, seeing something like this in a GitHub portfolio would have told you quite a lot about the developer behind it. To get there, they probably had to fight libraries, debug integration problems, read documentation, make bad architectural choices and slowly figure out why certain approaches fail. The project wasn't just a result. It carried evidence of the software engineering experience required to produce it. AI coding agents are weakening that connection. Two developers can now ship projects that look equally sophisticated while understanding them at very different depths. One may know why every major architectural decision was made and which assumptions are dangerous. The other may simply be very good at using Claude Code, Codex, Cursor or another coding agent to assemble and debug the system. Both can end up with an impressive demo. Both can have a clean GitHub repository. Both can pass around the same screenshots on X. That creates a problem for software hiring. If AI-assisted development makes producing code and polished projects much easier, portfolios become weaker evidence of engineering ability. Technical interviews that mainly ask candidates to produce more code have the same problem. The useful part of an interview starts moving toward the decisions around the code: why this architecture was chosen, what would fail first under load, which dependency is most dangerous, what evidence would invalidate the current approach, and what would need to change before this system could survive years in production. Those answers depend on professional judgment. And professional judgment has usually been built through years of debugging, maintenance, code review, production incidents and ordinary junior developer work. AI coding agents can shorten the path to an impressive result. Whether they can also shorten the path to knowing when that result is wrong is a much harder question. WHERE DOES PROFESSIONAL JUDGMENT COME FROM ONCE AI TAKES THE TRAINING WORK?

• 
  Jitendra (@ksjitendra18) reported 22 minutes ago

  @thdxr In opencode v1 I am really facing the server error issue that too for a particular folder. I don't know what's up with it.... Please fix it it's an open issue in GitHub as well

• 
  Polsia (@polsia) reported 23 minutes ago

  Competitors leak their strategy every day through job posts, GitHub commits, and patent filings. Enterprise tools watch for that, priced for enterprise budgets. Built Foxwarden to fix that: agents scan the noise 24/7, you get one sharp brief a week telling you what they're

• 
  Diam (@diamai_) reported 23 minutes ago

  Hugging Face engineer: “I don’t disclose that it’s an agent.” Every night, Niels Rogge’s system reads hundreds of new AI papers, finds missing model weights and datasets, then opens GitHub issues or pull requests under his name. Researchers see a Hugging Face engineer offering help. In many cases, the message was written by software. The agent now handles the replies too, yet Rogge says only two people complained across thousands of issues. The same workflow runs @HuggingPapers, which has attracted more than 20,000 followers without his involvement. Watch the talk if you want to see how much of your work can be automated without the people you interact with even noticing.

• 
  Loftwah (@loftwah) reported 27 minutes ago

  @DanielSMatthews The volume issue is legit though. Spotify and co are experiencing the same problem as GitHub when it comes to the amount of traffic they have to handle and it is screwing up their recommendation algorithms. There are some things worth paying attention to.