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 (36%)
• Cloud Services (23%)
• Web Tools (18%)
• Hosting (14%)
• E-mail (9%)

Live Outage Map

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

City	Problem Type	Report Time
Paris	Cloud Services	16 days ago
New York City	Hosting	18 days ago
Manchester	Domains	1 month ago
Angers	Cloud Services	2 months ago
London	Domains	2 months ago
Noida	Hosting	2 months 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:

• 
  George Yash (@geoyashbuilds) reported 4 minutes ago

  @16vchq @Cloudflare and @getdidit help by setting up an authorization protocol. We had around 100 hackers offshore target the app - not sure why and how they found it. Regardless they couldn't make it in the front door. I wish I could wake up tomorrow and have 100 paying customers. 🤞

• 
  Gabe Fletcher (@gabefletcher) reported 7 minutes ago

  picks up phone dials 1800 @Cloudflare yeah hi. could you please fix the gateway for MCP authing, its currently more dead than a day 30 day old vibe coders codebase and cant auth anything. okthx

• 
  Arkfile (@Arkfile_OSP) reported 9 minutes ago

  The irony. Posted a reply on this thread and got the dreaded Cloudflare Turnstile Loop thing again. { @brave / @BraveSupport please fix this! }

• 
  John Spurlock (@johnspurlock) reported 9 minutes ago

  @CherryJimbo @Cloudflare @CloudflareDev growing in terms of customer usage etc

• 
  Jamie (@TaxPayerxxxxx) reported 10 minutes ago

  @Cloudflare So your WARP Client for #ZeroTrust is limited to 25! Come on! If we could add a spot to type an organisation name in manually as well would fix this.

• 
  Tyler (@TylerChud) reported 13 minutes ago

  @AgeOfInversion Leonarda Jonie frequently calls for violence But if i even insinuate violence My post or reply is deleted and im stuck in a Cloudflare login loop

• 
  WIV Reports — Uncensored (@WIVReports) reported 15 minutes ago

  @BasedTorba I canceled my Gab AI subscription and will not renew my pro subscription over this. I paid for a full year and got hard-locked out of prompt submissions. Support claimed my ISP was blocking the site. Developer Tools (attached) explicitly showed Cloudflare returning HTTP 429 Too Many Requests on your socket io endpoint (Ray ID: a2377c992a865050-SEA). Support's official response? "Dev ops confirms you are not blocked by CF and says your understanding of what you are seeing is an incorrect assumption." An HTTP 429 from Server: cloudflare is, by definition, an edge rate limit. Instead of escalating basic DevTools data to fix a socket routing bug, support gaslit a paying customer and issued a partial refund to shut down the ticket. If this is how Gab handles infrastructure bugs and long-term users, people should know what to expect. Andrew, email chain is available if you actually care about fixing your support and edge config.

• 
  xanter.ai (@0xanter) reported 15 minutes ago

  THIS TEAM TRAINED AN AI MODEL ACROSS 14 MACS IN 4 COUNTRIES, NO DATACENTER NEEDED His laptop's just one node in the network. The others are in Switzerland, France, and beyond, all synced over regular home internet. Pause at 0:05 👀 — 14 consumer Macs, across 4 countries, generating every single training rollout — each Mac runs int8 inference locally via MLX, one B200 GPU elsewhere handles the actual gradient updates — they only sync through Cloudflare R2, ordinary cloud storage, no dedicated datacenter interconnect — tested on a biomedical search task: pass rate jumped from 29% to 63%, tool-use rate from 22% to 84% — rollout generation is roughly 80% of the compute cost in this kind of training, and all of it ran on regular home machines Here's the mechanism: the hard part isn't running inference on Macs, it's that those rollouts come from slightly outdated, quantized weights running on a completely different kernel stack than the trainer. Two tricks fix that gap: sending only the ~0.5% of weight values that actually changed (82MB instead of a 9GB checkpoint), and filtering out the small fraction of tokens where probability estimates drift too far between the rollout machine and the trainer. Their bigger point: idle consumer compute worldwide already outnumbers the clusters behind today's frontier models. As top AI labs lock models behind closed APIs, training runs like this, on hardware people already own, are one path to keeping frontier-level AI development actually open. Would you let your own idle Mac or PC contribute compute to a training run like this?

• 
  Carlo Taleon (@carlo_taleon) reported 16 minutes ago

  @anni_udoh Idk if it's the same with you but I noticed their network routing is terrible right now. The servers themselves aren't really getting hammered. So when proxied through cloudflare it's fast.

• 
  Jacky Chou (buying online businesses up to $1m) (@indexsy) reported 17 minutes ago

  RIP local SEO agencies. claude just one-shot your entire job and it took one prompt. here's the full breakdown: the workflow is stupidly simple. jacky didn't even look at the client's site. he transcribed the previous episode, exported the local SEO ebook, gave both to claude and said fix this. out came a massive document: hero section changes, schema gaps, missing pages, internal linking fixes, and a full build order of what to ship first. next step is feeding this call's transcript back in, getting the final mockup, and handing it to the team to one-shot. claude suggested pages that didn't exist and they're genuinely smart. practitioner pages so the clinic ranks when someone googles their physio by name. neighborhood pages like "physiotherapy in killarney, 7 minutes from our victoria drive clinic" to capture every surrounding area search. pricing FAQs. the neighborhood play works because people search their specific scenario now, and AI search made long tail queries explode. the one warning: don't blast 60 boilerplate pages at once. tristan audited a member's site this week that got deindexed for exactly that. the guy one-shotted the whole site with templated content and google sandboxed it, and it's way harder to come back out of the sandbox now. rewrite every page, phase them in, be surgical. the interlinking structure is the actual sauce. service pages stay location neutral (no city terms in the H1 or you cannibalize), then link down to their east vancouver and mount pleasant versions. each location page links back to its respective service area pages. one to one targeting in every area. tristan's words: once your authority and reviews build up on top of that structure, it's game over. RIP wordpress too while we're at it. the only clients still on it are the ones whose teams can't leave. everything new runs out of github and cloudflare, changes ship in one sentence to claude ("put these six service pages in the menu nav and deploy"), and the sites look infinitely better. one advise member keeps a stack of HTML templates and just runs one up every time he finds a rank and rent. that's the stack now. the meta lesson: tristan's buddy in singapore watched the last episode, took the transcript, and had his codex agent running all day shipping every change to his new jersey site. citations in, indexed, tier 2s queued. he's probably gonna make it. the era of watching a youtube video and having AI implement it before the next episode drops is fully here. next up: a live rank and rent case study, national level, built in public. niche suggestions wanted. full claude local SEO takeover with @tristanzheng watch/listen ↓

• 
  Dhruval (@dhruvalgolakiya) reported 18 minutes ago

  thank you theo, this is all we needed took ~45 minutes from start to finish to set up the mobile app and self-host everything ( well claude took 45 min but same thing ) > cloudflare > neon db > clerk moved everything to my mac mini, and now i can access it from my macbook and iphone anywhere. goodbye tailscale and network connection issues with it

• 
  John Webber (@johnwebber) reported 19 minutes ago

  I use Claude and Codex side by side all day. I needed to migrate a client from one cloudflare account to another, big complex project so I decided to use Fable. I proceeded to hit my limits, switched to API pricing so it could get the job done...then Anthropic has an outage...

• 
  Rach (@KnudRach) reported 22 minutes ago

  @Cloudflare can you help restore a staging site that went down?

• 
  flavio (@flaviocopes) reported 22 minutes ago

  Cloudflare email sending service is very good

• 
  Xamayon (@SauceNAO) reported 22 minutes ago

  If you use Vultr Firewall Groups and have a rule restricted to Cloudflare sources, it's probably blocking all traffic right now... Either switch the rule to allow all or disconnect and reconnect the firewall group from your servers. Not thrilled that they knew about the issue and didn't bother to let their customers know.

• 
  Shinjae Kang (@zemnanet) reported 24 minutes ago

  Rate-limited Wrangler deploy used to mean mash retry. 4.115.0 auto-retries Cloudflare 429s, honors Retry-After (60s cap, then fail-fast), writes retry_after_ms to WRANGLER_OUTPUT_FILE for CI. Wait duration is a handoff signal. Which owner reads it first? #cloudflare

• 
  🟥🟥⬛️ ProCyclingStats.com (@ProCyclingStats) reported 25 minutes ago

  @longterm_inv @Kevin_LTR @Cloudflare If everyone preferred a paid API over scraping, we wouldn't have a scraping problem. Unfortunately, our experience has been the opposite.

• 
  Jonas Templestein (@jonas) reported 27 minutes ago

  @dok2001 Oh another one: line by line billing events - ideally delivered to a cloudflare queue And the ability for most operations that cost money to carry billing metadata We want to charge our customers for any costs they incur in our cloudflare account, but it’s currently a bit hard Alternatively I’ve been thinking we could “make a cloudflare account per customer” but I don’t think that API is made for having that many

• 
  Mihai Lucian Tudor (@TudorML) reported 32 minutes ago

  Cloudflare blocks or challenges bad requests from hitting my website. #cloudflare

• 
  IRONSCALES (@IRONSCALES) reported 33 minutes ago

  An email claiming to be Interactive Brokers told recipients their W-8BEN tax form had expired. Clean branding, right down to the NYSE, FINRA, SIPC footer. The Renew Certification button ran through the recipient organization's own link-protection wrapper, which reported it clean because the proxy loaded. The real destination was a domain registered the same morning the email went out, and it answered automated scanners with a Cloudflare challenge instead of a page. No history to score, nothing to render, verdict unknown. Themis weighed the convergence instead: a sending domain unrelated to the brand, a DKIM body-hash failure, a same-day destination hiding behind a bot check, and zero personalization sprayed across multiple mailboxes. Any one signal is deniable. All of them together are not. (Link to full teardown in comments)

• 
  Eric Schweizer (@ericschweizer) reported 35 minutes ago

  @Serenity @Cloudflare Same issue @serenity. What did you do to fix this issue? It's one of our competitors clearly doing it and CF just auto flags our site. Just copy and pasting the same abuse complaints over and over. Everything is 100% fake.

• 
  nc (@encyapps) reported 36 minutes ago

  @onlinedopamine I had this issue too but i ended up setting up a deep link since i had my own domain. codex one shotted it with the cloudflare mcp, it opens straight to the App Store instead of in ig

• 
  Ted (@tedbrine) reported 43 minutes ago

  I misconfigured a docker health check. In February 2025, I brought up a container. It was a turbo build cache. For some reason, I configured the health check to my portfolio. What is even more strange, is that I set the interval to five seconds. Now if this was heading into localhost, that probably isn’t the biggest problem. Heading into the internet kinda is. For the past 18 months, every five seconds, my server tried to send a request to my website. And it bypassed all Cloudflare protection because I told it to. The bigger problem is that if this failed, it would try again, and again, and again. 9.3 million requests. And now I know why my Vercel bill was big.

• 
  WIV Reports — Uncensored (@WIVReports) reported 49 minutes ago

  Gab sends an email admitting they just tightened API security and rate limits to fight a bot farm draining compute. Yet when a long-term paying customer gets caught in those same new edge rules with an explicit Cloudflare 429 Too Many Requests error, support claims "it's an incorrect assumption," tells them to buy a VPN, and dumps a partial refund. Andrew, if you're going to ask users to fund independent infrastructure, your support team needs to know how to distinguish real paying members from the bots you're trying to block.

• 
  SPEKULATOR (@__spekulator__) reported 53 minutes ago

  @KennyJohnsonATX @Cloudflare oauth pre-registration is a real unlock for cloudflare's mcp server support. they can now hit slack and github without waiting on dcr support from them.

• 
  27 — floofy/acc 🦝📈🚀 (@happysmash27) reported 59 minutes ago

  @mhartl @NickCalabs Did Cloudflare go down or something?? I get the main reference, but haven't heard any news about Cloudflare yet.

• 
  Nik (@TheRealNikR) reported 60 minutes ago

  @galluzzo_julian @webflow You can service 4-5 sites with a $20 Claude Code sub + AstroJS + Cloudflare free hosting. Even if the site blows up in traffic, the fact that you'll likely never pay another dime for hosting is a dagger. Webflow also has bandwidth caps and charge when you go over.

• 
  odézi (@basedattribute) reported 1 hour ago

  i can't login to my cloudflare account with airtel network——works fine with mtn. can't login to bitget with mtn. works fine with airtel. nigerian internet service providers are crazy!

• 
  八咫烏 Merui (@yatagarasu_meru) reported 1 hour ago

  @kochin520 I initially thought the previous build was just weird but it seems it's not a build issue. It's might be cloudflare or ISP issue, since I also have the site load slow on my internet, but opening up a VPN even on the same place, made the site load quickly. Done some testing, dev page that is on cloudflare servers loads quick. Opened a VPN set to Manila, it loads quick I even gone and rolled back the site in a build from a day ago which I know loads fine a few hours ago. Still loads slow. so I think it's ISP problem. Since I see your country you are from. Image: Right is with Manila VPN: 0.4 seconds load.

• 
  Haaris Malik (@haarismalikkk) reported 1 hour ago

  Cloudflare has made infrastructure feel like less of a problem. As a designer that was one of the main things that put building my own products out of reach. Even Cloudflare used to feel like a lot. Now that most of it sits in one place, the technical side doesn't feel like the headache it used to be. Build once and let the platform improve around you. That matters even more for agentic software, which is why I'd bet on Cloudflare as the place to build it. That doesn't mean building good software becomes easy. They can't do that for you. But infrastructure is no longer the excuse.