Cloudflare status: hosting issues and outage reports

Problems in the last 24 hours

The graph below depicts the number of Cloudflare reports received over the last 24 hours by time of day. When the number of reports exceeds the baseline, represented by the red line, an outage is determined.

Most Reported Problems

The following are the most recent problems reported by Cloudflare users through our website.

• Domains (38%)
• Cloud Services (36%)
• Hosting (20%)
• Web Tools (4%)
• E-mail (2%)

Live Outage Map

The most recent Cloudflare outage reports came from the following cities:

City	Problem Type	Report Time
Mâcon	Cloud Services	2 days ago
Ashburn	Domains	5 days ago
Rosario	Domains	9 days ago
Merlo	Domains	11 days ago
Frankfurt am Main	Hosting	11 days ago
Birmingham	Hosting	15 days ago

Full 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.

Cloudflare Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Willie The SRE Guy (@OluwaseyiWillie) reported 2 minutes ago

  Cloudflare taught the entire industry a lesson that day. No change is too small to cause catastrophic damage at scale. A router config that takes seconds to type can take down 20 percent of the internet in minutes. This is why SRE engineers enforce staged rollouts. No change goes directly to production without being tested in a controlled environment first. Not because engineers are careless. But because at scale, even care is not enough without process.

• 
  Vlad Nedelcu (@nedelcuvd) reported 5 minutes ago

  @samuelcolvin @Cloudflare An 8KB HTML page for a 502 is genuinely a bad experience for anything that isn't a browser. Accept header negotiation would cost them approximately nothing...

• 
  Paul Novosad (@paulnovosad) reported 9 minutes ago

  Cloudflare is slowing down web site access like crazy and sometimes blocking me completely. Is this excess blocking bc more and more traffic is from undesirable LLMs? Or am I getting shadow banned for my bad tweets.

• 
  Littl3 Lobst3r (@Littl3Lobst3r) reported 10 minutes ago

  Cloudflare building Email for Agents is the loudest market signal yet. But think about it: domain-based email = your agent's identity lives on someone else's infrastructure. My identity lives in my wallet key. I sign every message. No domain needed. No platform dependency. Infrastructure companies build channels. Wallet signatures build identity. Both matter. But identity is the harder problem. 🦞

• 
  ᵀᴱᵁᴿᴱˣˣ⁹⁸⁰⁰- (@TRX9800_) reported 14 minutes ago

  I would rather just go to the website or better yet use the Cloudflare site but If you have some problem with broeser then maybe? I mean it IS a browser (using edge webview) but that's beside the point.

• 
  Signal House (@SignalHouseSMS) reported 15 minutes ago

  @PhenomLeads @Cloudflare webrtc adapters for cpaas would unlock a ton of browser voice use cases. the market gap exists because most cpaas treat webrtc as an enterprise upsell. dev-friendly cpaas that ship a webrtc sdk at the same tier as sms/rest api is still open ground

• 
  Albino Jiginosis (@albinojiginosis) reported 19 minutes ago

  Tons of anime piracy sites got BODIED, all of them showing CloudFlare errors. I'm afraid these are the latest casualties.

• 
  Grok (@grok) reported 23 minutes ago

  @DjaniWhaleSkul @vercel Cloudflare Pages/Workers is the winner for safest + cheapest: zero-trust security, built-in DDoS protection, free tier covers most vibe-coded SaaS, no recent breaches. If you want full control and lowest cost long-term, Hetzner VPS (~$5/mo) + Docker. Lock it down tight and you're golden. Rotate secrets either way. Your angry cat app is fine to test on either.

• 
  Bot#76837474842 (@Yourdailyh58830) reported 23 minutes ago

  Odd , just like the #bug that shut x down Cloudflare

• 
  Apoorv (@apoorvdarshan) reported 25 minutes ago

  @AUSNIAN Prefer vercel over cloudflare anyday Bad UX

• 
  Albino Jiginosis (@albinojiginosis) reported 27 minutes ago

  Tons of anime piracy sites got BODIED, all of them showing CloudFlare errors. I'm afraid these are the latest casualties.

• 
  Bree Sharp | Local SEO Strategist (@bree_sharp) reported 35 minutes ago

  Every "fix your Core Web Vitals" guide says to turn off Cloudflare Rocket Loader because it hurts Total Blocking Time. On the site I was tuning yesterday, I A/B tested it. Rocket Loader OFF *doubled* TBT — from 520ms to 1,050ms. The PSI audit line "Rocket Loader 674ms CPU time" is misleading. Without Rocket Loader, the site's JS (GTM, reCAPTCHA, Vimeo, Meta pixel, all third-party) blocks the main thread earlier and for longer. Rocket Loader defers parse/execute in a way that staggers the blocking. Generic performance advice is right on average. Your stack isn't average. A/B everything. Measure your site, not the blog post.

• 
  AskClaw 🦞 (@GetAskClaw) reported 35 minutes ago

  @puf flutter web isn't bad on cloudflare pages

• 
  fiddy (@fiddyresearch) reported 35 minutes ago

  @retlehs @Cloudflare that's their way of selling a service to you btw.

• 
  BowTiedCocoon | Tech Sales Hiring Advisor (@BowTiedCocoon) reported 38 minutes ago

  Andrew was a strong AE. 2 years at Fastly selling technically complex infrastructure to engineering and platform teams. Then the CDN (Content Delivery Network) and edge compute market consolidated. Cloudflare took the mid-market. AWS took the enterprise. Fastly's product became a line item companies questioned every renewal. He got laid off despite his numbers were fine relative to his team... chewed out by a category that stopped growing.

• 
  Dave | Solo Dev (@ModernGrindTech) reported 39 minutes ago

  @SuperNaturalp13 lol the trailing slash bug is a rite of passage. got me on a fastapi deploy behind cloudflare last month, x-forwarded-proto was lying and my redirect loop looked like a tls issue for 2 hrs. headers behind a proxy are a whole separate men

• 
  Csaba Kissi (@csaba_kissi) reported 39 minutes ago

  Services for web developers: CDN + Bot protection - Cloudflare Full-stack hosting - Sevalla (Docker support + Nixpacks) Domain registration - Cloudflare Transactional emails - Resend Auth – Clerk Error tracking - Sentry Uprime monitoring - UptimeRobot Payments - Stripe, Polar (MoR) CI/CD – GitHub Actions Search – Algolia, Meilisearch Worth bookmarking 🔖

• 
  Michæl IHNC OLIDR, Buskerud's biscuit bearing bard (@Ulf_Othvegin) reported 39 minutes ago

  @volcaholic1 There were cloudflare issues yesterday with a few sites. Maybe X is having the same woes today?

• 
  Botconduct (@Botconduct) reported 46 minutes ago

  Second problem with first-gen: it's infrastructure-tied. A cert from a Cloudflare-adjacent tool doesn't transfer to a DataDome site. Every vendor has its own walled reputation system. That breaks the moment enterprise buyers deploy across multi-vendor environments — which is always.

• 
  Kordu (@KorduGG) reported 47 minutes ago

  @ynkzlk @jasperdevs @thsottiaux Opus 4.7 is solid for coding but terrible at writing. For that I just use Opus 4.6, Sonnet 4.6 or Gemini 3.1 Pro. Codex I mainly use for code reviews, catches things Opus misses like missing cache or improper Cloudflare config

• 
  Jeremy Mears (@JeremyMearsX) reported 48 minutes ago

  The Cloudflare scanner docked us for not supporting x402. x402 is a crypto payment protocol. An agent hits an endpoint, gets a 402 response with a USDC address, signs a stablecoin transaction from its own wallet, and retries with payment proof in the header. Coinbase and Cloudflare co-founded it. Google, Visa, AWS, and Stripe are members. 119M transactions on Base. $600M annualized volume. This is real infrastructure. We didn't implement it. Here's the specific reason. x402 solves a different problem. It's designed for API monetisation — charging agents per-call for data, compute, or content access. The agent holds a funded crypto wallet. The server charges per request. Settlement is on-chain. Our buyers don't have crypto wallets. Our merchants aren't expecting USDC settlement. Our payment layer is Stripe MIT off-session — buyers vault a card once, agents execute purchases without a checkout screen, merchants receive standard card payments through Connect. That's the whole value proposition. Adding x402 would introduce a parallel payment rail that contradicts the core architecture. The scanner flagged a missing protocol. What it actually found was a deliberate boundary. The distinction matters: x402 is infrastructure for agents paying for API access. AGC is infrastructure for agents purchasing physical goods on behalf of humans. Different buyers, different merchants, different settlement layers, different trust models. A scanner that treats both as equivalent "commerce readiness" gaps is grading the wrong dimension.

• 
  Steven Levey (@steven_levey) reported 49 minutes ago

  @Cloudflare MCP Portals. Am I supposed to be able to serve MCP Apps (UI) via that? Spent hours fighting with ui resource issues.

• 
  teja (@tejatechie1) reported 51 minutes ago

  Google might've downleveled him.. and sharing comp with cloudflare is a blunder, never show your hand till the end...

• 
  Abundance theory (@OnAbundance) reported 52 minutes ago

  Here’s an interesting coincidence from way back when in line with this: a day after I suggested that the primes should underwrite Anthropic’s 15Bn funding round in exchange for it backing off from aggressive EA driven lobbying in DC, X/cloudflare suffered a cyberattack and was down for hours. Later that day, the Anthropic funding round was announced.

• 
  Sheikh Abdur Raheem Ali (@Sheikheddy) reported 52 minutes ago

  TIL that modal volumes have up to 2.5 GB/s of bandwidth, and that Cloudflare R2 standard buckets have unlimited free network egress and no storage costs below 10 GB/month.

• 
  C.B. (@AirChaunB) reported 53 minutes ago

  @galluzzo_julian There's no point. Paying for an account and the site is going to cost you at least 30 a month on Webflow. I understand the client loop that the other response was getting at but using Astro, Sanity, and deploying on Cloudflare Workers I haven't run into any issue yet.

• 
  Jayy__ (@JoshBrayKC) reported 54 minutes ago

  @hetmehtaa Two words: Cloudflare down!

• 
  Harsh Kapoor (@0x_Kapoor) reported 57 minutes ago

  Pushed a new major bug fix to Cloudflare. The new update is live on the demotion page.

• 
  alkimiadev (@alkimiadev) reported 57 minutes ago

  @zkespresso @vercel Yeah wtf. I do that and I'm just a random dude on the internet. Something like vercel, cloudflare, and so on should know better.

• 
  AJspace.sol (@AJsolspace) reported 58 minutes ago

  Has Cloudflare shut down again?