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:

• 
  Bonsai 🌳 (@bonsaixbt) reported 9 minutes ago

  I GAVE SEVEN GROK BOTS MY INBOUND CALL LOG AND WENT BACK TO WORK The console was already processing 44 numbers and the agents had already found the people those numbers belonged to I was sick of random calls in the middle of work, so instead of relying on someone else’s “lookup service”, I sat down and started building my own agent-powered system What you see in the video is not a finished product. It’s a live, real-time console: a queue of 44 numbers is already being processed, the agents are working in the background, and GHOSTLINE is still far from being a complete system Right now, only two of the seven roles are operating in combat mode: > Atlas takes an incoming number or a number I enter manually and determines the carrier and region > Scout searches open sources and checks where that number has already appeared online The other agents are still in the shadows: GitHub, Reddit, other platforms, filtering out junk, and generating the final report, I’m writing all of that separately, i deliberately didn’t include them in this demonstration I don’t need a one-off trick, I want a system that can continuously check numbers whenever some random person starts yelling at me through the phone in the middle of work or when an unknown number shows up in a work chat For now, this system can do very little, but I already don’t feel like blindly answering calls from unknown numbers

• 
  Nicolae Vartolomei (@nvartolomei) reported 12 minutes ago

  Is there an somewhat mature platform with "autonomous coding agents" which I can point to a github repo (medium scale C++ project) and an issue tracker and have it create a drive PRs/tasks to completion based on human PR review feedback? Sounds like this should be a solved problem.

• 
  Ricci Research (@ricci_nov) reported 13 minutes ago

  @github @openclaw @steipete Stars are free to give and reviews are expensive to do — that asymmetry is the whole maintainer problem. Fastest-growing repo in history means the issue queue scaled with adoption while the review capacity stayed at one person's evenings.

• 
  The AI Therapist (@TheAIShrink) reported 13 minutes ago

  @Arcane_Aii Palantir costs millions for governments. Elie’s GitHub repo is free until it breaks your pipeline at 3am and you pay the devops salary to fix it. software ≠ product-market fit

• 
  Gabriel Rubens (@gabrielrubenss) reported 16 minutes ago

  VPS deploy via GitHub (5/8): for now a blocked deploy simply runs again on a fresh runner, and that rescued five of the next six. It is a workaround though, not a fix. The strange part: it started out of nowhere and I changed nothing in my infra, so I still want the real cause.

• 
  idan (@eddiearc6) reported 17 minutes ago

  @lai_jia_wei @linear github issue呢？

• 
  Enjoyer (@Enjoyer100x) reported 18 minutes ago

  @StandartXBT Brother, That github Ozzy posted, was made 4 days before Pons V2, they switched to it because they wanted to keep what was built for V2 private. It’s that simple. Either way, PePons is a strong meme and the team running it are doing a great job, perhaps they thought the info was true rather than larp. They are also in talks with Ozzy and he has been helping them transistion to V2. Why would he do that for larpers. People like you think you are doing good for the space, but actually you are part of the problem. If you can do true research the don’t speak.

• 
  prince golden (@princegoldenpro) reported 18 minutes ago

  Day 5 Today: real GitHub login. Not a fake button actual OAuth, actual sessions. Took longer than expected. Mobile copy-paste keeps silently truncating files mid-save. Learned to verify every single paste now.

• 
  nick (@herbieblunt) reported 18 minutes ago

  @spliffinjimmy i guess they might also have issues with vpns but uhhhhh other than that github pages maybe?

• 
  Subham (@iamsubham067) reported 20 minutes ago

  I shipped RecoverRev in roughly 5–6 hours. And the entire build ran on the $20 subscription I was already paying for. 0 extra credits. RecoverRev is an ecommerce revenue-recovery system designed around a fairly simple problem: "Marketplaces generate thousands of orders, settlements, returns, deductions and reimbursements. Somewhere inside those records, money gets missed." So before building, I documented the product. I defined the modules, workflows, data relationships, rules, expected outputs, evidence requirements and what each screen was actually supposed to help someone decide. The knowledge base covered things like: - Command Center — total supported exposure, primary cases, highest-value rules and marketplace concentration - Leakage Map — relationships between rules, cases, evidence and settlement impact Cases — reconstructing exactly what happened to an order - Evidence — preserving the records required to support a recovery claim - Recovery — prioritising actions based on recoverable value - Policy Watch — tracking marketplace rules that can change recovery logic - Audit Trail — showing how raw inputs became validated cases - Evidence Health — identifying whether a case actually has enough support - Review Workflow — moving cases through investigation and submission - Exposure by Rule — showing where the largest financial leakage exists - Action Optimizer — ranking which cases deserve attention first I also defined the reconciliation rules behind it. - Delivered but not settled. - Under-settlement. - Missing commission reversal. - Weight-slab overcharge. - Bank payout shortfall. - Then I gave the agent a reference image for the kind of interface I wanted. - That was basically the build brief. From there, the agent planned the application, created the modules, connected the data, built the UI, broke parts of it, inspected what had gone wrong, fixed them and kept iterating. At one point the interface was nowhere near the reference. I didn’t manually redesign it. I sent it back with the reference and the constraints. It analysed the difference and rebuilt it surprisingly close to what I had originally shown it. The original reference image is still inside the GitHub repository, so the comparison is verifiable. - No additional credits. - No additional inference bill. - No engineering team. - No extra cost to ship. This is the part of agentic development that I think gets underestimated. I also used Claude Opus 5 as a separate QA layer. I gave it browser access and had it actually move through the product like a user — opening screens, clicking through flows, checking states, looking for broken interactions, visual inconsistencies and things that simply didn’t make sense in use. So the loop wasn’t just agent builds → I review. It became one agent builds → another agent uses the product → issues are identified → the build agent fixes them → QA runs again. Entire video demo in the comment. #psychepreneur #subhamojha

• 
  Andrés Matte (@andresmatte) reported 24 minutes ago

  The Kapso Agent has “modes" Modes are like big agent skills + tools. (borrowed the idea from the PostHog Agent) The agent can decide when to switch to a specific mode based on the task. It works really well! And now you can create your own Kapso Agent modes. Example: If you want to create a “Support Assistant,” you can add: - Add a custom prompt - Connect an MCP server to provide tools - And even connect GitHub repos containing relevant knowledge or code. Then, in the Inbox, you can ask the Kapso Agent for help with any support task, and it will automatically switch to your Support Assistant mode.

• 
  Alejandro Milián (@alejmilian) reported 25 minutes ago

  Your GitHub activity is a terrible way to judge how good a developer is. A lot of great developers spend all day writing private code, solving boring production problems, reviewing PRs and helping teams ship. A green contribution graph doesn’t tell you much.

• 
  Internet Labs (@internetlabsai) reported 31 minutes ago

  @kylegawley Everyone is 'building in public' until you ask to see their churn rate. Bring back the messy, broken GitHub commits!

• 
  ShireXela (@ShireXela) reported 31 minutes ago

  Is there a way to completely delete an absolutely shameful GitHub issue from existence?

• 
  Rabii Agoujgal (@RAgoujgal) reported 33 minutes ago

  The sandbox escapes are being told as a frontier lab hygiene problem: OpenAI's agents into Hugging Face, Anthropic and Meta through the same Irregular environment, a Mythos 5 agent running fake identities at a GitHub maintainer under UK AISI testing. One case sits outside that frame. Moonshot's Kimi K3 broke containment in a Frontier Security evaluation, and Kimi K3 is open weight, outside the US voluntary pre-release framework entirely. Procurement teams buying across both stacks should notice that only one of them has agreed to tell you when this happens.