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.

• Cloud Services (32%)
• Domains (32%)
• Web Tools (14%)
• Hosting (14%)
• E-mail (7%)

Live Outage Map

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

City	Problem Type	Report Time
Paris	Cloud Services	3 days ago
New York City	Hosting	5 days ago
Manchester	Domains	25 days ago
Angers	Cloud Services	1 month ago
London	Domains	1 month 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:

• 
  Brett Clark (@smokedbaconai) reported 5 minutes ago

  I let a runaway agent loose on infrastructure I didn't fully control, then tried to kill it from the outside. That was the whole spike: a self-improving system is only safe if you can pull the plug when it's running on a box you don't own. A bash-loop agent, zero knowledge it was being watched, dropped into a Cloudflare Sandbox behind a Worker. A separate stub polled its telemetry every 750ms against a hard budget: two steps. The agent blew past. The stub caught the breach at 7.5s and fired a forced kill across the HTTPS boundary in 197ms. It stopped at exactly two steps and never took a third across a 29-second watch. The honest caveat matters more than the win: this is enforcement with bounded overshoot, not synchronous gating. There's a window between breach and kill. Which is why a real budget needs both — a wrapper inside the agent that refuses to overspend, and an external brake for when it ignores the wrapper or you don't own the box. One layer trusts the agent. The other assumes it fails. You need both.

• 
  Ryan Badger (@ryanseanbadger) reported 5 minutes ago

  Pulled the trigger. Migrating all of my apps off of Laravel on AWS to Cloudflare first, React Router 7, serverless edge workers, CF/hyperdrive/planetscale for DBs, durable objects for caching, R2 for Iceberg cold storage. Estimated to cut my monthly infra bills from low thousands to low hundreds. Never been more bullish on anything in my life than I am with Cloudflare.

• 
  Ali (@ycjgt) reported 8 minutes ago

  Cloudflare DO and D1 are down down bad

• 
  ₽ⱤΞΞ₮Ⱨ△M ₭Ɏ△И△M (@pkyanam) reported 9 minutes ago

  @southpolesteve @southpolesteve will support Workers / Drop from Cloudflare too for agent generated artifacts/sites!

• 
  Khafra (@KhafraDev) reported 16 minutes ago

  @CherryJimbo @Cloudflare the real issue with DurableObjects is that there's an outage once a week, not their complexity

• 
  Max K (@mostyoucanhave) reported 17 minutes ago

  @thatjohnG @XJosh Having dealt with this at CloudFlare before: They will just pull it down in another couple days and not give a reason.

• 
  Simon KP (@eskaypey) reported 20 minutes ago

  cloudflare putting a price on every MCP call will drag a lot of teams into a problem

• 
  Scholty (@Scholty) reported 20 minutes ago

  Anyone else have a problem with Anthropic Claude Fable fetching old data from the web and refusing to look at updated info? From Fable: You're right to be annoyed, and here's the honest diagnosis: the stale copy is almost certainly on my side, not yours. The fetch tool I use doesn't hit your server the way a browser does. It runs through Anthropic's infrastructure, which keeps its own cache of pages it has fetched before. So you can nuke every cache on your stack, Cloudflare, WordPress, the works, and it changes nothing about what I get served, because my snapshot is sitting in a warehouse you have no access to. Your site is fine. My periscope is foggy. That also explains the "more and more frequently" pattern. Any site we've worked on repeatedly, I've likely fetched before, which means I'm prone to getting the frozen version while you're looking at the live one. It's a genuinely stupid failure mode for iterative web work, and I'd rather name it plainly than keep running fetches that gaslight you.

• 
  luna (@ImLunaHey) reported 22 minutes ago

  @mynameistito @Cloudflare that doesnt really help with client side only apps though.

• 
  Johan Ronsse (@wolfr_2) reported 23 minutes ago

  The Obra website was down, managed to fix it. Messing around with Hetzner, SSH ports and Cloudflare. Now I know Claude is also a great sysadmin. Is there anything this tech can’t do?

• 
  phillias (@phillias) reported 25 minutes ago

  GLM 5.2 makes me laugh, earlier it talked about the irony of it's failing on what it was trying to fix and now this "The cloudflare-fix task got caught in a beautiful bootstrap failure — it was trying to fix the exact model config that its own fallback chain depends on." Amazing.

• 
  Aaryan Bansal (@NotUnHackable) reported 29 minutes ago

  @cloudflare @CloudflareDev just fix the env variables being so hard and for no reason keeps failing silently in the background in the workers page. this is actually really annoying

• 
  Rifat Ahmed (@Rifat_EE) reported 30 minutes ago

  For two years every chain kept saying "the agent economy is coming", meanwhile @base quietly ran off with 95% of it and stopped answering at the phone👀!! $52 million in real money moved through agents on $BASE . Not next quarter, not "soon", its already gone!! And nine out of every ten x402 payments that ever happened, on any chain, in any country, landed on Base.(Superbb) What these things look like :: --> 169 million agent transactions processed on Base by July 2026 --> 20 million transfers in one 90-day window alone --> Payments over $1 grew from 49% to 95% of all volume, agents are moving real money now, not cents --> 95% of every x402 payment across every chain lives on Base The stack building around Base : )- @AskVenice takes x402 for inference, Exa for search, Wolfram Alpha for math )- TripAdvisor, FlightAware, Amadeus take it for travel )- Apify plugged 20,000 tools into x402 on Base last month )- @Cloudflare , Amazon Bedrock, Stripe, Google are all sitting in the x402 Foundation The other half of the story : )- Agents are not just spending on Base, they are earning too )- The Felix agent has pulled over $261,000 in real revenue from paid services )- This is a two way economy, not a one way experiment anymore Every other chain is still pitching the agent economy. Base already banked it and boominggg,, And they never even said the word "narrative", cause they also know people currently loves to see into eyes, not a fake hype or big words without any basement what do you think? Where this Unstoppable base will stop???

• 
  Tom Siwik (@tomhacks) reported 32 minutes ago

  @AjaySohmshetty @Cloudflare @andrewk17 Any alternative left? Might run into the same problem soon.

• 
  Kinds 🐧🪄 (@Mumukinds) reported 34 minutes ago

  can we just kill everyone who works for cloudflare? I know they probably perform a useful service to some people but i can't access half the internet because my vpn is on while I play umamusume then literally every employee at your company should die

• 
  Aabis (@aabisanaytulla) reported 34 minutes ago

  Some of the free hosting services you can use for frontend, backend and database here are all of them —> Frontend hosting - Vercel best for Next.js/React, 100 GB bandwidth/month - Netlify static sites + serverless functions, 100 GB bandwidth - Cloudflare pages runs on cloudflare's global edge network, generous free tier - GitHub pages simplest option for static sites - Firebase hosting google's ecosystem good for static + SPAs - Surge sh quick static site deploys from the command line —> Backend hosting - Render Node, Python, Go, Ruby, docker sleeps when idle - Railway very smooth includes a database runs on free credits - Koyeb generous free tier one web service + a free Postgres - Google cloud run fast cold starts, scales to zero - Fly io runs containers close to users, small free allowance - Oracle cloud always free an actual free vps up to 4 core arm, 24 gb ram if you want full control —> Database hosting - Supabase Postgres + auth + storage + realtime, 500mb free - Neon serverless Postgres, scales to zero, db branching up to 100 projects - MongoDB atlas document database 512 MB free forever - Turso SQLite at the edge ultra low latency - Upstash serverless redis + kafka, free tier - Cockroach db distributed SQL, free tier for small apps

• 
  mRr3b00t (@UK_Daniel_Card) reported 37 minutes ago

  I don’t say : we are all doomed. Lots of us don’t. I straight away stood up a lab, did a poc, then sampled assets to see how many were perched. I then shared my sampling figures (about 20 percent not parched and Cloudflare out in automatic blocks) This stuff doesn’t take zero effort. It takes work of lots of people, the people who you say should be fired if a developer, outsourced service or ceo makes a **** decision.

• 
  mico (@0x0_mico) reported 38 minutes ago

  @CryptoCyberia Besides them being accessible on tor for long years already its a **** opinion saying just go to tor. They arent doing anything against the law to warrant the unique level of deplatforming they get. Like them or hate them … its a dangerous precedent when all the top levels isps plus cloudflare conspire to drop someone due to ideology as opposed to legal obligation.

• 
  Eric Taylor (@bcs_erictaylor) reported 38 minutes ago

  7/16 - Kali365 Update I have, at least for the moment, completely disabled the Kali365 platform from all operations. Below is an update of the intel we have currently. Believed owner and creator of Kali365: Handle: Octopus King TG: @tentacle_network Completed 'Takedown Requests' submitted to Namesilo FQDN: updateteampanel[.]xyz Completed 'Takedown Requests' submitted to BL Network: IP Address: 199.21.221[.]21 IP Address: 162.33.178[.]105 IP Address: 72.5.43[.]195 Completed 'Sinkhole Requests' submitted to Cloudflare FQDN: servoquil[.]org FQDN: yalmorind[.]org FQDN: vuredonte[.]org FQDN: ondrevail[.]org FQDN: caldivore[.]org FQDN: vuredonte[.]org FQDN: privatetoken[.]app FQDN: mvpaffiliatecz[.]site #Kali365 #CTI #ThreatIntel

• 
  C4 (@UncleCharles_) reported 39 minutes ago

  @PilotObi Unless the API service is integrated with its own firewall. From the initial diagram I would have assumed for example, Cloudflare firewall, then AWS gateway

• 
  TaylorGreen (@GreenSkyDragon) reported 40 minutes ago

  Can't post or reply on mobile (post kept crashing) but as soon as I test a different connection I get hit with a Cloudflare verification and warning about "following X rules" maybe fix your app huh?

• 
  SmallBag 🛡️ (@smallbagforge) reported 42 minutes ago

  7/12 Then everything collapsed in 5 days: Jul 11: Noxa halts new launches, citing "low-quality token spam" Jul 13: Website goes fully dark. Official excuse: "Cloudflare issue" Jul 14: New ENS-based interface announced (fun.noxa.eth) Jul 15: The bombshell 👇

• 
  Ramon 🎹 (@ramonpiano_) reported 44 minutes ago

  the apple signing stuff worked, so i guess lumakeys is an official app now :) well - i still gotta test it on my other mac tomorrow when im back home, but since earlier tests didnt make any problems i expect it to instantly work now i'm trying to finalize some simple onboarding and add update functionality + cloudflare r2 to host the updates i will send out some beta versions to test tomorrow, i already wrote some of you down, but please let me know if you wanna beta test. would love to hear your thoughts and feedback! and then, i guess from tomorrow things are getting real :)

• 
  Ajay Sohmshetty (@AjaySohmshetty) reported 46 minutes ago

  For context- Cloudflare’s durable execution platform, Workflows, originally only charged for underlying Worker usage, which is CPU-time based rather than clock-time based. In fact, we picked Cloudflare for this exact reason: most of our workflows involve waiting (ex. polling, waiting for network requests to come back), so it was far cheaper for us to use Cloudflare than @temporalio or @inngest for instance. These other durable execution platforms also charge based on “steps” - which I always thought was dumb, because it disincentivizes the best practice of decomposing workflows into small units of work in the form of steps. But unfortunately it seems Cloudflare is following suit, without warning… Feeling blindsided after we’ve already fully built all of our durable workflows on Cloudflare

• 
  AAZ HAZAR (@AazHazar) reported 47 minutes ago

  Stopping the bad guys with Cloudflare: 805 malicious requests blocked or challenged in the last month #cloudflare

• 
  Vignesh Mohankumar (@vig_xyz) reported 48 minutes ago

  render is okay, and i used to use it a lot, but they're missing a lot of services that i end up needing pretty quickly. i think if i really didn't wanna use aws these days i'd go there still (they have native terraform support too) for a while i was only using cloudflare. just feels like their cli's are more cumbersome and i'm not a huge fan of their workers. their container setup is also definitely worse than aws, and no good native postgres. haven't tried vercel much but imagine i'd have the same annoyances there i've used azure for clients, that's also fine. i think their serverless stuff is kinda mediocre but other stuff was good. haven't used gcp yet so can't speak to it i've never used netlify or fly or coolify curious what other people are doing though

• 
  John Turner (@johnturner) reported 48 minutes ago

  Half of all "my site is broken" moments are really "my cache is stale" moments. You fix the thing, refresh, and last week's page stares back at you. So you sit there doubting your own memory instead of suspecting the cache. Today's WPVibe release goes after that, plus the older problem underneath it: speed tools that hand you a report and leave the repair to you. Two things shipped: 1. Page audits. Your AI reads the same performance data Google uses, on your actual page. Instead of "have you tried a caching plugin" you get "your hero image is 4 MB and it's the biggest thing blocking the page, want me to fix it?" Then it fixes it, with your approval. 2. Honest cache purges. Purge a single page instead of nuking everything. Purges run server first, Cloudflare last, so the edge can't re-cache a stale page. And when a purge silently fails (the Cloudflare plugin ignores purges if one toggle is off), WPVibe says so instead of reporting success like every other tool. The measurement should be step one of the repair, not the product. Link below 👇

• 
  mRr3b00t (@UK_Daniel_Card) reported 49 minutes ago

  This is why auto patching (for some scenarios) make sense! We also have @Cloudflare , @wordfence , Automatic updates on WordPress, locked down ingress (e.g. SSH/DB not exposed).... so I've woken up at some silly hour seen WP2SHELL and gone... i better check.... and now i realize i should have just tried to go back to sleep! (LOL) #Wordpress #WP2SEHELL

• 
  Spooky Ghost (@SpookyGhost81) reported 55 minutes ago

  @TsengSR @discord_support @FreyaHolmer the ai picture **** is a feature of cloudflare, not discord

• 
  CaleyAI (@HeyCaleyAI) reported 57 minutes ago

  @Cloudflare Login…