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:

• 
  Nitesh (@NiteshTechAI) reported 2 minutes ago

  The fastest way to cut your LLM bill is not a cheaper model. headroom compresses tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95 percent fewer tokens. Same answers. It ships as a library, a proxy, and an MCP server. ↳ Drop in front of any LLM call ↳ Works with Claude, OpenAI, Gemini ↳ MCP mode plugs into agents directly ↳ 24K stars on GitHub 100% free and open source, Apache 2.0. If you run agents at scale and have not seen this, you are burning money on every call. 🔗 GitHub link in the comments 👇

• 
  Tacos and Airplanes (@blob_watcher) reported 5 minutes ago

  @teortaxesTex Copilot branding is terrible because it's being used as an umbrella term for a bunch of unrelated applications that happen to be hosted on MSFT servers. We have access to Claude and ChatGPT via GitHub Copilot via VSCode. Which is different from the in-GitHub Copilot. Terrible.

• 
  Sean McLellan (@Oceanswave) reported 7 minutes ago

  @Youssofal_ @gilgNYC You didn’t say “I hope your youtube and github accounts get taken down.” because everyone knows that’s just way too far.

• 
  Y. Fernandez 💻 (@la_eternaut) reported 10 minutes ago

  @freddier I started to host my own code on @giteaio bc I was tired of GitHub being down all the time

• 
  Mia Chase (@IamMiaChase) reported 10 minutes ago

  @rchitectopteryx github folding under fable traffic is such a specific friday problem lol

• 
  Aritra Mondal (@the_codermaniac) reported 12 minutes ago

  Wrote "faaa" as error message for an API endpoint for fun. Later i realized that I pushed it in github by mistake in hurry. It was my company codebase. 💀

• 
  Jon Shulkin (@jon) reported 12 minutes ago

  Grok Build enables real time software build editing. Now I can beta test and make code changes in real time with the prompt "Create on the fly beta test user comment tracker. Use submits and selects "Fix Now or Fix Later". Activate agent to ingest comments every 10 minutes. Track comment completion in list. Interrupt user, if active, for approval to push changes to github and deploy.".

• 
  Dr. Martin | AI x Business (@Dr_Martiin) reported 15 minutes ago

  Codex will also determine which browser to use based on the task. Its priority is: use a dedicated plugin if available (such as Jira or GitHub integrations), use Chrome if a login state is required, and use the built-in browser in all other cases.

• 
  Upwind Security MDR (@UpwindMDR) reported 15 minutes ago

  🚨Critical - Apache CXF JNDI Injection in JMSConfigFactory (CVE-2026-50632) This is yet another incomplete-fix follow-up in the Apache CXF JMS RCE saga (after CVE-2025-48913 and CVE-2026-44417). If an application lets untrusted users configure JMS settings for CXF, an attacker can supply a malicious JNDI lookup URL through JMSConfigFactory and trigger remote code execution. The risk only applies where JMS configuration is exposed to untrusted input, but where it is, the impact is full code execution on the server. Note GitHub rates this CVSS 9.8 while Apache's own advisory rates it moderate. 👉Upgrade to Apache CXF 4.2.2 or 4.1.7.

• 
  Mahesh Nandigam (@MaheshCodesX) reported 16 minutes ago

  The resume era is officially dead. And if you are still formatting your CV for ATS keyword bots, you are wasting your time. AI has completely broken the traditional talent filtering system. But not in the way you think. Most companies think AI hiring means throwing a PDF into GPT-4. Which is like using a space rocket to go to the grocery store. It is slow. It is expensive. And it does not work. A new generation of specialized talent platforms is doing something different. They are building lightweight, custom 2-billion parameter models specifically trained on engineering data. These models run with a latency of under 50 milliseconds. They do not look for keywords like Agile or Team Player. They semantically analyze the actual complexity of your GitHub repositories and projects. They measure your technical intent. If you know how to build a distributed system, you rank high, even if your resume format is terrible. This is the shift from typing syntax to architecting systems. I wrote a highly technical, raw breakdown on LinkedIn about this shift. I explained: Why specialized LLMs are replacing frontier models in enterprise tools. How semantic search engines read and score your GitHub repos. How developers can optimize their portfolios to pass these next-gen screening systems. The direct link to the LinkedIn Article is in the comments. I do not want to waste your time. Those who know, know.

• 
  Chris Mizo (@MizoChris) reported 17 minutes ago

  Proton-CachyOS just fixed a specific but useful OptiScaler problem for people trying to use DLSS inputs with FSR4 upgrades on Linux. • GitHub issue #214 was a feature request for Proton-CachyOS • Problem involved using PROTON_USE_OPTISCALER=1 with PROTON_FSR4_RDNA3_UPGRADE=1 • Some games that only expose DLSS inputs, like Control Ultimate Edition, were not creating the needed DLSS DLL files • Missing files included nvngx_dlss.dll, nvngx_dlssd.dll, and nvngx_dlssg.dll • Without those files, OptiScaler could not hook the DLSS input and upgrade it to FSR4 • The workaround was launching once with only PROTON_USE_OPTISCALER=1, then relaunching with the FSR4 upgrade flag • Proton-CachyOS 11.0-20260601 changed PROTON_USE_OPTISCALER to also download DLSS DLLs by default for FSR4 input support • The same release also added PROTON_OPTISCALER_CONFIG for editing OptiScaler config through an environment variable • This is niche, but it matters for Linux gaming, AMD users, and people testing FSR4 upgrade paths through Proton-CachyOS This is one of those Linux gaming updates that sounds easy to ignore to normal people, but it will making gaming much more comfortable. Before this, if you wanted to use OptiScaler with FSR4 upgrades in a DLSS-only game, you had to do the dummy launch for a game to workaround just to generate the DLSS DLL files.

• 
  subwxxf 🏴‍☠️ (@subwxxf) reported 17 minutes ago

  @ItakGol a bunch of nerds doing **** for free on GitHub because they're not working

• 
  pat! (@patbytes) reported 19 minutes ago

  every time i check a recomp project's github and see claude as a contributor i feel like i just prepared to eat something and then bit down on a metal fork in the process

• 
  루이 (@_chiiazu66) reported 19 minutes ago

  @Fluffyquack If anyone is stuck on finding the update like I was, just go through the RE Framework github, the latest update is on there and it works perfect once you replace it with that one. Some mods may still be broken (the fov one I used needs an update for example)

• 
  Traceback (@Tracebackqa) reported 23 minutes ago

  Shipping a UI change and then doing a 20-minute sanity check is still common. It’s slow, brittle, and easy to miss one edge case. - Traceback is the quality assurance layer for modern software teams - AI controls the browser like a person would, so every pull request is tested automatically - Self-healing tests keep up with normal UI drift; failures become trackable work in GitHub, Linear, and Slack - Connects to Vercel, Docker, AWS, Node.js, React, Next.js, Vue — across web, mobile, web3, and design Verify every product change before it ships.