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	14 days ago
New York City	Hosting	16 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:

• 
  Hunter Guo (@hunterguo101) reported 2 minutes ago

  Anthropic just shipped the biggest MCP update ever (MCP 2026-07-28). Most takes are protocol trivia — stateless, extensions, auth. The real signal, if you actually build: your distribution channel just got upgraded. MCP is now the HTTP of the AI era — and the pipe is enterprise-grade, runs on serverless, and can render your UI inside Claude itself. The numbers: 400M+ monthly SDK downloads (4x this year), TypeScript and Python each past 1B cumulative, 950+ MCP servers in the Claude store used by millions daily. The standard war is over. The question is whether you're on the rail. Three things that matter: — Core went stateless. Deploy your server to Lambda/Vercel/Cloudflare Workers and scale horizontally behind a plain load balancer. Building an MCP server now costs about as much as writing one serverless function. — Extensions are first-class. MCP Apps render interactive UI right inside the Claude chat (sandboxed iframe). Your interface lives inside Claude — you don't build a front-end, pull users, or educate a market. Make the capability; the channel carries the UI and the users. — The pipe reaches inside the firewall. MCP Tunnels connect Claude to internal servers with no public IP and no inbound firewall rules. Figma, Intuit, Netlify, Zoom are on board. The takeaway: stop building another app to fight for attention. Build an MCP server and let the agent ecosystem be your distribution. Agents are the new distribution channel. This update paved the road.

• 
  yohance (@yohanc33) reported 5 minutes ago

  @leinwand @eastdakota @Cloudflare Cool, I happened to apply for this fall’s SWE position (and the previous fall, as well as the previous summer, no luck so far :p). Think I could get a referral? I’ve used several cloudflare infra products (R1, Workers, Workers AI, Workflows, etc). Working on mastercard’s service mesh this summer using Go.

• 
  Ride The Leaders (@ridetheleaders) reported 6 minutes ago

  @RichardMoglen $NET has one of the strongest weekly bases in $IGV Inference has to live near the user and Cloudflare already owns a 300 city network. The weekly backs it up with a six month base, RS line at new highs ahead of price, handle forming 9% off highs. Undercut and reclaimed the 21ema today while $IGV broke. Earnings in 10 days should resolve the base...

• 
  Bojan Josifoski (@twistedroute) reported 8 minutes ago

  @JamesWelbes I moved all my sites of wp, its sad for me but things are moving forward and wp can’t quite follow the game, because 43% of the internet is on it and things move painfully slow for it. Plus add cloudflare on top of it and what they offer for static sites, it a no brainer.

• 
  eddiebro (@EddiebroRon) reported 9 minutes ago

  I decided to write an iOS shortcut that will encrypt a message before sending it over iMessage and then decrypt it by only the person that has the correct 256 bit key, read it and then reply to me using the same process BUT I’m using the same type of Double Ratchet protocol Signal uses so that every message has a different encryption key and if someone were to gain access to one of the participants private key, they would not be able to do anything with it as it changes each message. Anyway, I am in the process of auditing my shortcut because iOS 27 adds features to use AI to write your shortcut for you and then also auditing and I was writing the audit and so excited that I was within five minutes of testing all of this that I just spent hours doing, and I go to open the door and walk inside with the dogs and then look down at the Shih Tzu and she look like a Christmas tree with cockleburs as decorations and I wanted to cry. It’s hard to believe that I’ve actually pulled more off of her in the past, but this was still pretty bad. And now I shall go test my shortcut yay. If it works as intended I might put it up in the gallery and then I will post a link here. The method to get a private key bound to someone else (using a temp key they provide to you from the same page) is a website that I threw up real quick on Cloudflare and then the key and all processing is done on device using JavaScript so technically anyone could use it without risk of me or anyone else being capable of decrypting any messages 🤷

• 
  Theo - t3.gg (@theo) reported 12 minutes ago

  "npx t3 connect" This one's been a lot of work. You can now set up remote control for T3 Code on any internet-connected box with literally one command. All for free. T3 Connect is a minimal open source tunnel layer allowing you to control T3 Code instances remotely without needing Tailscale set up. I've been daily driving it for a month and it has changed how I code. Julius and I put a lot of effort into making setup as smooth as possible. Step 1: Install Claude Code, Codex, OpenCode, or Grok Build Step 2: Run "npx t3 connect" Step 3: Click link and sign in Step 4: You can now control that computer on T3 Code web, desktop or mobile (dropping very soon) We are currently providing this for free (s/o CloudFlare for bumping our tunnel limits). Every user can connect up to 3 devices. We don't want to charge for this, but if the bill gets unacceptable we may have to change course. If you hit limits or have issues, you can always fork, self host, or use Tailscale. T3 Connect may seem like a small ergonomic win. Tbh that's exactly what it is. Regardless, it's one I'm really proud of.

• 
  Velko Minchev (@VelkoMinchev) reported 15 minutes ago

  Day 29 📅 We had our first outage. Cloudflare ran maintenance on their infrastructure and TimerOS went down with it for a stretch. Not our code — but that hardly matters to someone who can't log in. What I care about is that everything behaved: → Incident alerts hit my inbox immediately, so I knew before anyone had to tell me → The status page reflected it live — anyone checking got a straight answer instead of a spinner → And I improved it mid-incident: incidents now open in the TimerOS support system too, so there's a written trail, not just a red dot You don't find out whether your transparency is real until something breaks. Ours held. Today: last stretch of road, heading home 🚗 Back at the desk tomorrow — Cloudflare and I have unfinished business. 😤

• 
  Pete Stewart (@pj_stew) reported 15 minutes ago

  @founder_sohan You can use mist open source models with Worker AI, which is a fraction of the cost of SaaS services like OpenAI, and if you use Cloudflare AU Gateway with it, you can also use any of the main SaaS services and switch between models as needed. You can see in the article I wrote yesterday about video transcription that I used Whisper on Worker AI to make a much cheaper transcription service for Achene.

• 
  🜑 (@1casie) reported 16 minutes ago

  > be me > macbook air is bricked, needs a reinstall > no second mac to build a bootable installer > find a github actions workflow that builds macOS installers for free, no mac required > perfect.jpg > trigger a build, artifact is a 17.6GB .dmg.img > click download > kb/s > literal kb/s > theoretical connection is 10MB/s > discover artifact lives on azure blob, signed URL expires in 10 min > write a bash script that resolves the azure SAS url fresh, hands it to aria2c with 16 parallel streams > 16 connections, still kb/s > mfw it's a per-IP throttle, not per-connection > repo author literally has a line in the workflow: "Enable Cloudflare WARP for faster download" > install Cloudflare WARP on my ubuntu box > single stream jumps 20KB/s -> 1.6MB/s > aria2c 16x -> 10MB/s, pinned to my ceiling > 17.6GB in ~an hour, sha256 matches github's published digest > oknowweiscooking.webp > dd it to a USB SSD > paranoid it won't boot apple silicon because it was built on an intel runner > boots apple silicon just fine, i was wrong, shut up > installer runs, reinstall chugs along > STOPS > "OS Personalization" PREFLIGHT_PERSONALIZE portioncomplete:0.15000 estimatedtimeremaining:-1 > 70000 log lines of stalled:NO > the mac can ping apple/com > it can ping cupertino > but personalization "no connections witnessed" > i don't ******* get it > try to sniff the wifi to see what it's talking to > can't > WPA2 isn't a hub, it's a switch, i can't decrypt the mac's traffic as a third client > feel stupid > dig into it > test apple's personalization signing servers from my box > gs/apple/com -> 000 > gsp-ssl/apple/com -> 000 > albert/apple/com -> 403 (works) > gs/apple/com is the signing server and it's IPv4-ONLY > my ISP (DIGI, romania) has broken IPv4 routing to apple's 17.0.0.0/8 > ICMP works, IPv6 works, CDN works, the ONE /8 macOS needs to sign firmware is a black hole > that's why it could "ping cupertino" but personalization silently died > fix: repoint the mac's IPv4 default gateway from the router to my ubuntu pc > my pc has IP forwarding + NAT masquerade + Cloudflare WARP > mac's IPv4 now tunnels through cloudflare, bypassing my ISP's broken path > watch the signing handshake complete in tcpdump, SYN -> SYN-ACK -> ClientHello -> kilobytes of signed data > portioncomplete climbs past 0.15 > weactuallydidit.jpg > STOPS AGAIN at 0.29 > silent on IPv4 > oh > the mac's IPv6 still goes straight to the router, bypassing my box > i literally cannot see it > disable IPv6 on the mac so everything falls back to the working IPv4 path > it reboots > flashes the apple logo 4 times > i'm dying > it boots > it works > 29 hours > i had to MITM my own home network because my ISP couldn't route to one (1) /8 > the mac is back > i'm buying a framework >text still isn't green

• 
  adam (@theCTO) reported 19 minutes ago

  never, we’re on @convex and @railway and @cloudflare everywhere, all at once

• 
  Rach (@KnudRach) reported 23 minutes ago

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

• 
  Mitch Flindell (@buildo_baggins) reported 24 minutes ago

  @meaganrgamache Using Think for a project, had to switch to openrouter because glm inference on cloudflare was very slow

• 
  2024=2023+1 (e/acc) (@commondoubts) reported 27 minutes ago

  @paularambles Site is down now. @Cloudflare to the rescue?

• 
  Citedy (@citedycom) reported 31 minutes ago

  Cloudflare Dashboard is temporarily unavailable. Claude down. What next. Aliens?

• 
  Nashon (@WEGNERJOASH) reported 33 minutes ago

  @Okwachjamal Generating 15 gb of data takes time. Had an option of running it on a vps but don't want to login to my cloudflare through that. Writing it to local sqlite is faster and ensures that I can resum sending it to remote db if I have a network issue. I can also test locally.

• 
  Techtitanmb (@techtitanmb) reported 35 minutes ago

  Cloudflare is fixing network-performance problems in Istanbul that could affect connected websites and services. Are you noticing unusual loading issues? #techtitanmb

• 
  sam with his computer (@samwcomputer) reported 38 minutes ago

  cloudflare just made blocking ai training bots the default on ad-supported pages starting september. opt-in security was never going to work, defaults are the only setting that actually gets used

• 
  DR.PIXEL (@pixelpatchit) reported 38 minutes ago

  Swapdesk Alpha: DNS Expansion and Reliability Improvements As Swapdesk Alpha continues to mature, we plan to expand beyond our current Cloudflare and Google DNS support by adding additional DNS provider options. These providers have given us a strong foundation during the alpha phase, allowing us to focus on building and testing core functionality while maintaining reliable connectivity. During the THORChain swap process, users may occasionally experience intermittent connection issues while the application is retrieving required swap information, such as the deposit address needed to complete a transaction. These issues are not consistent and do not occur during every swap, but when they happen, they can create confusion because they occur during an important step in the transaction flow. It is important to clarify that an error message during this stage does not automatically mean a transaction has failed or that funds are at risk. In many cases, the issue is related to communication between the application and supporting services, which may include DNS resolution, API availability, network routing, or other connectivity factors. To improve reliability, future versions of Swapdesk Alpha will introduce broader DNS support, improved fallback handling, and additional resilience measures. The goal is to provide users with a smoother experience, reduce unnecessary interruptions, and ensure that Swapdesk remains dependable as adoption grows. We appreciate the feedback from early users helping identify areas where reliability and user experience can continue to improve

• 
  Huintellimance (@Huintellimance) reported 41 minutes ago

  MCP just shipped its biggest update since launch. The protocol is now fully stateless — and that changes everything for how AI agents connect to tools. Here's what happened: The Model Context Protocol dropped spec 2026-07-28 yesterday. Since late 2024, MCP has gone from "spec nobody heard of" to nearly half a billion SDK downloads per month. Both the TypeScript and Python SDKs crossed 1 billion total downloads. It's the de facto standard for how AI models talk to external tools, databases, and APIs. But until now, MCP had a problem: it was stateful. Every MCP connection required a handshake (the initialize/initialized exchange), a persistent session ID, and an open bidirectional stream. That meant your MCP server had to live on a long-running process — no Lambda, no Cloudflare Workers, no edge functions. You needed a VPS or a container that could hold state across requests. That's over. The new spec removes the session handshake entirely. Every request is now self-describing — it carries its protocol version, client identity, and capabilities inline. No session ID. No persistent connection. Any HTTP request can land on any server instance behind a plain round-robin load balancer. What this unlocks: Serverless MCP servers. Deploy to Lambda, Workers, or any function-as-a-service platform. Pay per request, not per idle server. Edge deployment. Put your MCP server at the edge, close to the user. Cold starts are irrelevant when there's no session to establish. True horizontal scaling. No shared state means no sticky sessions, no Redis, no session stores. Just spin up more instances. Built-in observability. Method and tool names now travel in HTTP headers (Mcp-Method, Mcp-Name), so gateways can route and authorize without parsing the body. But the protocol didn't just strip things away — it added smart replacements. The old server-to-client requests (like asking the user for confirmation mid-tool-call) used held-open streams. Now they use Multi Round-Trip Requests (MRTR) — the server returns a structured "I need input" response, the client gathers it, and sends a follow-up. All stateless. All over plain HTTP. Tool catalogs now carry cache hints and deterministic ordering, so clients can cache them and keep upstream prompt caches stable across reconnects. No more re-fetching everything on every connection. There's also a formal extensions framework now — Tasks for long-running work, MCP Apps for server-rendered UIs, and Enterprise Managed Authorization for orgs that need it. Why this matters: MCP went from "cool protocol for local dev tools" to "production-ready infrastructure for the entire AI agent ecosystem" in one release. The deployment story is now as simple as any REST API. If you've been holding off on building MCP servers because the infrastructure overhead felt like too much — that excuse is gone. TypeScript, Python, Go, and C# SDKs are all updated. Migration notes are live. The breaking changes have a 12-month deprecation window. What MCP servers are you building — or planning to build — now that serverless is on the table?

• 
  ObsidianIntel 𐎡𐎼𐎠𐎴 🇮🇷 (@IntelObsidian) reported 42 minutes ago

  @TMobile Most importantly, why did it take half a day before you acknowledged it? Learn from Cloudflare, they go down nearly daily, but at least they are quick to announce it, then they are detailed in their after action report.

• 
  Informer |-/ (@_Nformer) reported 45 minutes ago

  I love how half the comments are just people who get salty when Cloudflare goes down and the other half are mad about Cloudflare having a monoply.

• 
  Nick Blow (@NickBlow) reported 46 minutes ago

  @ejc3 @rivet_gg Nothing, I used it a lot at a previous company, it just doesn’t generalize nearly to azure or google cloud or on prem. I know there’s some work towards this, but it suffers the same problems as Cloudflare really.

• 
  Mr. Code NJ (@perpetualtalk) reported 48 minutes ago

  @saltyAom Does Cloudflare support bun + elyisa? Is this in the rust rewrite?

• 
  احمد (@born2code) reported 49 minutes ago

  @threepointone I did that with fargate (so I can use OpenAI subscriptions), sqs and dynamo. The problem with workers is that cloudflare blocks the OpenAI subscription calls (I suspect this this OpenAI buy blocked at CF edge), which is a non starter at the rate I use tokens

• 
  Efikcoin Eternal (@EfikcoinEternal) reported 51 minutes ago

  Stopping the bad guys with Cloudflare: 525 malicious requests blocked or challenged in the last month #cloudflare

• 
  Jeromy Sonne (@JeromySonne) reported 56 minutes ago

  Slack is failing, some web apps are throwing me cloudflare 504s. Incoming cloudflare outage in 3, 2, 1

• 
  Polymarket Alpha (@Polymarketalpha) reported 57 minutes ago

  🚨 JUST IN: Bots now generate more web traffic than humans. According to Cloudflare, bots accounted for 57.5% of global web page requests in June 2026, while human traffic fell to 42.5%. This marks a historic turning point for the internet. AI crawlers, automated systems and intelligent agents are no longer a minor part of the web—they are becoming its dominant users. Cloudflare’s CEO had previously predicted that bot traffic would not surpass human traffic until 2027. Instead, the milestone arrived a year early. The internet is rapidly shifting from a network built primarily for humans to one increasingly accessed, indexed and operated by machines. Technological progress—or the beginning of a less human internet?

• 
  Jacob Hands (@jachands) reported 58 minutes ago

  @dok2001 Fix the Cloudflare codemode MCP server to properly request scopes for all APIs - I had to stop using it because it's missing auth for so many APIs.

• 
  Shigaraki Tomura (@grimmjow_sats) reported 59 minutes ago

  NVIDIA, MICROSOFT, IBM, CLOUDFLARE AND CROWDSTRIKE JUST FORMED A 37-ORG ALLIANCE — AND THEY'RE NOT AFRAID OF THE MODEL. they call it the Open Secure AI Alliance. it shipped an open-source stack, NOOA, to test, trace, audit and govern how agents actually behave in production. read what they're guarding against and it's obvious: → not "the AI goes rogue" → over-permissioned agents with access they never needed → tool calls nobody can trace after the fact the scary part of an agent was never its IQ. it's the keys you handed it and the logs you don't have. 37 of the biggest names in compute and security just admitted the same thing at once. everyone shipping agents is one untraced tool call away from the incident. follow and bookmark before your stack becomes the case study

• 
  Rahul Gangotri (@rahul__gangotri) reported 1 hour ago

  `@ supabase/supabase-js can never hibernate on cloudflare DO as it require the changing ref. only way forward is to support both. costly, non hibernating supabase client package and my own package which is hibernate friendly and costs way less. lets do it!