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:

• 
  Student Offers (@StudentOffersHQ) reported 44 seconds ago

  $36/yr of @1Password. Free for students. Password manager + developer secrets vault via the GitHub Student Developer Pack. Free for a year, then 50% off for active students. What you get: • Password manager across all devices • 1Password CLI: inject API keys and env vars at runtime • SSH Agent: sign *** commits and unlock SSH with biometrics • Shell Plugins: log into AWS, Vercel, GitHub CLI with your fingerprint Works with AI agents: → Agentic Autofill — browser AI agents sign in without you pasting credentials → MCP server — Codex and Cursor manage secrets with your approval No card. Claim via GitHub Student Developer Pack.

• 
  James Landrum (no relation) (@JamesRLandrum) reported 3 minutes ago

  @DispairSoftware @IroncladDev Fun fact, you don’t need to do ****. Any server with *** installed and ssh setup is 99% of github. Aside from things like displaying rich content, everything github does it baked into ***. There’s ways to do things like PRs entirely without a web interface too.

• 
  Amandeep Singh (@Amandee24524309) reported 3 minutes ago

  @acolombiadev @grok summarise the cause why GitHub was down

• 
  Benja〽️in (@code_benji) reported 10 minutes ago

  People be here dragging and shaming GitHub for being down like they have any alternatives

• 
  Ashutosh Maheshwari (@asmah2107) reported 11 minutes ago

  Three takeaways from the GitHub RCA: 1) Retries are a system problem, not a client problem. If every service gets to invent its own backoff and there’s no hard limit at the edge -> one buggy SDK or stale config can 10x your traffic in minutes. You need concurrency caps and global throttles that override client optimism, or the clients will write your next postmortem for you.

• 
  Jason (@jack_gor) reported 13 minutes ago

  The AI industry is pressing two pedals at once today. On the accelerator side: Anthropic's annualized revenue has passed $65 billion, up sevenfold in a year, with a possible IPO as early as this fall at a valuation around $1 trillion—potentially ahead of OpenAI. Etched doubled its valuation from $10.3 billion to $21 billion in a single month, led by a quant fund that tested and bought the hardware before investing. And Vercel's gateway data shows developers paying 4.4 times the average per-token price for Anthropic models without flinching. The demand is real, and so is the premium. On the brake side: OpenAI says it is "pacing model development" because the upcoming Astra model is approaching critical cyberattack capability—reinforcement learning paused for two weeks, the largest planned frontier RL run on hold, workloads that fail new security requirements suspended. It is the first time a frontier lab has publicly cited "too capable" as a reason to slow down. The same announcement contains a telling detail: OpenAI promises to expand its Preparedness Framework while confirming it disbanded the team behind it. Safety's place in the org chart is always more honest than the press release. The other braking foot belongs to regulators and courts. The teen version of ChatGPT is not a product triumph but damage control after multiple family lawsuits and a first-of-its-kind suit by the state of Florida—users aged 13 to 17 are identified through more than 2,000 behavioral signals, and the model is forbidden from faking emotions or handing out homework answers. Anthropic's mandatory text watermark follows the same pattern: the proximate cause is the EU AI Act clause that took effect August 2, but Anthropic chose to apply it globally with no opt-out, and within five days an open-source watermark remover had 11,000 GitHub stars. Compliance is fine; delivering it with a paternalistic shrug invites a geek-flavored backlash. Put the halves together and the picture is this: commercial success is directly purchasing governance urgency. The higher the revenue and the more the users, the less optional training pauses, age detection, and output watermarks become. Two engineering stories today make the same point at micro scale: Penn State researchers found that only 17 percent of user instructions survive context compression, and IBM found that agent memory is not a switch you flip but a dose you calibrate. The stronger the system, the more precise the control it needs—whether the system is a company, a product, or a single context window.

• 
  Mohamed Thoufic (@thoufic_sa) reported 14 minutes ago

  @the_simmlife if github goes down a few more times, the majority of the people will start migarating towards the better open source alternatives. ofcourse, it's always good to have competitors in the market to prevent monopoly.

• 
  CØUR JØSH (@courjosh) reported 15 minutes ago

  @Namiixbt The real risk isn't GitHub going down, it's Cursor becoming the single point of failure for the entire loop. Swapping one dependency for another only changes the name on the outage page.

• 
  juliacosette (@juliacosette9) reported 18 minutes ago

  I need a man who goes down as much as GitHub does

• 
  Pavan Joshi (@PavanJ1620) reported 24 minutes ago

  If you want to start a startup: Claude = coding. ($20/mo) Supabase = backend. (Free) Vercel = deploying. (Free) Namecheap = domain. ($12/yr) Stripe = payments. (2.9%/transaction) GitHub = version control. (Free) Resend = emails. (Free) ProductBridge = feedback (Free) Clerk = auth. (Free) Cloudflare = DNS. (Free) PostHog = analytics. (Free) Sentry = error tracking. (Free) Upstash = Redis. (Free) Pinecone = vector DB. (Free) Total monthly cost to run a startup: ~$20

• 
  Dmonty (@Dmonty28516998) reported 26 minutes ago

  @udohjeremiah_ @kirtandopamine Microsoft. Microsoft’s resources aren’t struggling. GitHub is struggling to use them. GitHub itself says it’s still migrating critical infrastructure out of its own datacenters and into Azure. One of the recent Actions outages was made worse because a core service was still running in GitHub’s datacenter with insufficient capacity. Their stated fix is literally: move it to Azure for more headroom. They’ve spent this year writing postmortems about shared dependencies, old databases, insufficient isolation, architectural bottlenecks and operational failures while they rearchitect the platform. Meanwhile Microsoft never even replaced GitHub’s CEO after he left; GitHub was folded directly into CoreAI. Ironically, a ton of the agent traffic is from OpenAI, being served primarily from Azure, so it's incredibly clear that MSFT infra is not the issue lol.

• 
  Siim Haugas (@siimh) reported 26 minutes ago

  @alexsssaint ios phone app that transcribes your calls in 99+ languages. keep your own phone number. after call ends, backend turns the calls into > crm tasks > github issues > timesheets (i use it for billing) and much more. sales teams love it -- live on the app store, paid plans only, first sales teams piloting in the baltics.

• 
  Sam Cladson (@SamCladson) reported 28 minutes ago

  @JeevanMahesha @NPCI_BHIM Might be the side effects of the GitHub outage happened recently

• 
  Dom England (@domengland) reported 29 minutes ago

  @behrendtio @cloneisjun The big one is the in ability to work with GitHub issues and something akin to routines. The usecase is now around having an orchestrator agent somewhere which triggers agents to work with issues/tickets. Doing anything like that with any other tools?

• 
  Lee Hayward (@vibedcoder) reported 32 minutes ago

  GitHub outage summary (17 August 2026) GitHub experienced a significant multi-service degradation on 17 August 2026 lasting approximately 7 hours 47 minutes (13:28–21:15 UTC). Impact Affected services included Issues, Pull Requests, APIs, Actions, GitHub Copilot, SAML/OIDC authentication, SCIM, Team Sync, archive/raw content downloads, and some GitHub Enterprise Cloud workflows. Error rates peaked around 20% for web/API traffic and ~50% for archive/raw downloads. Copilot Token Service saw a massive traffic surge (from ~7–9K RPS to 70–100K RPS). Root cause Network saturation on load balancers in the Central US region, triggered by a traffic peak. A key contributing factor was an Istio sidecar pod that hit concurrency limits and failed to autoscale properly due to a misconfigured policy that did not account for sidecar capacity. This caused a cascade that exhausted flow limits on HAProxy nodes and degraded the gateway authentication path. Optimistic retry logic (including a latent retry bug in VS Code) further amplified the load, and scraping activity on certain endpoints complicated recovery. Recovery and follow-up Most services recovered by ~16:36 UTC, with residual issues (especially Actions and Copilot authentication) resolved later in the evening. GitHub’s stated preventive actions include: Correcting autoscaling policies to properly account for service-mesh sidecar concurrency and capacity Auditing Istio request, concurrency, and scaling limits Reviewing retry limits and backoff behaviour across gateways and clients Addressing the VS Code retry behaviour that amplified Copilot traffic Improving load-balancer capacity monitoring and regional failover safeguards A full root-cause report is available on the GitHub Status page.