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 (39%)
• Cloud Services (29%)
• Web Tools (14%)
• Hosting (11%)
• E-mail (7%)

Live Outage Map

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

City	Problem Type	Report Time
Manchester	Domains	18 days ago
Angers	Cloud Services	30 days ago
London	Domains	1 month ago
Noida	Hosting	1 month ago
Jewar	E-mail	1 month ago
Braga	Web Tools	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:

• 
  Swai (@IAmSwaiDhanoa) reported 16 seconds ago

  @eastdakota Looked at this with my CISO! Works for billing and LLM gateway but the thing I didn’t get was “How do you help sales right size the model to task”. To @threepointone’s point..a user shouldn’t think about that. There’s components to support what Im describing in the Cloudflare platform. I’ll take a stab.

• 
  opossum in blossom 🏳️‍⚧️ трансгендерная мышь (@ok_im_merging) reported 30 seconds ago

  @nostalgiacore Cloudflare be like: we should build the entire encryption system on this ****

• 
  Subhankar (@subhankarP) reported 5 minutes ago

  Cloudflare just pushed AI crawlers to pay publishers for content. The old scrape-first era is ending. AI economy is starting to price inputs it used to treat as free.

• 
  Tom Rush (@tr4777) reported 7 minutes ago

  I’m activating KER-453 as the sole security child and using the OFF issue workflow to carry it through current-truth preflight, minimal remediation, adversarial review, one PR into staging, automatic staging deployment, bounded test-identity proof with cleanup, and Linear closeout. Before editing I’ll identify the exact profile-contract, persistence, authorization-consumer, OpenAPI, and regression-test files; production, provider, credential, Cloudflare/Render configuration, and unrelated auth semantics remain outside scope. /codex

• 
  Manfred Neustifter (@supermanfredX) reported 7 minutes ago

  @Cloudflare The worktree name (which is the issue name)

• 
  ZEE (@CodeWithZeee) reported 7 minutes ago

  @oljimenez @NathanFlurry I mean with companies like Cloudflare especially but also true for some AWS and GCP services, the tech and infra line is pretty blurred which adds to the conversation… take a message queue or distributed event bus for example you could argue you could augment the infra with the tech, use a library that’s backed by a specific platforms infra but in reality using something like googles pubsub over aws SQS isn’t going to be plug and play there’s a level of lock in regardless so it feels like fighting the fight is kinda pointless. the alternative is significant investment in abstraction which any good software engineer knows can be a curse if reached for too early into a product or companies life cycle. So yes lots of nuance but I still stand by it being something that isn’t as big as people make it out to be… dig in focus on solving the problem

• 
  Zaher 🦋@zaher.dev (@zaherg) reported 8 minutes ago

  hey @RhysSullivan , it seems the latest version of the selfhosted it broken (at least the docker image is) Cannot find module '@cloudflare/worker-bundler' from '/app/apps/host-selfhost/dist-server/serve.js' missing sourcemaps for /app/apps/host-selfhost/dist-server/serve.js

• 
  Javi Lopez ⛩️ (@javilopen) reported 9 minutes ago

  Summary: Here's the gist of what Pieter Levels (@levelsio) actually does. He stopped coding locally. Instead he rents a cheap $5/month Hetzner VPS, SSHes into it with Termius (an SSH app he uses on both his iPhone and MacBook, roughly 50/50), and installs Claude Code right on the server with npm install -g @anthropic-ai/claude-code. Then he just types claude and talks to it. He calls the whole approach "VibeOps." The mind-bending part is that Claude edits the code live on the production server. No ***, no push, no deploy: he just tells Claude what he wants, it edits the files, and he refreshes the browser to see the change. He runs it in a "bypass" mode (an alias that skips all the permission prompts) so it just goes, and reports he ships around 10x faster this way: he even says he "outran his todo list." In 12 months it only broke production twice, for about 10 seconds each. He makes it safe with a specific recipe: install Tailscale first, lock the Hetzner firewall so only Cloudflare can reach port 443 and only his Tailscale IP can reach SSH on 22, put Cloudflare in front, never allow password login, and then ask Claude itself to set up fail2ban and automatic updates. He's clear this is for hobby/solo projects, not sensitive stuff: for a real company he'd use a staging server. Because everything lives on the server, he can jump from laptop to phone without losing his place (one tmux session per site, auto-reconnecting) and literally codes from planes. The most impressive trick: to build a native iOS app, his VPS Claude Code SSHes into a rented Mac Mini (MacinCloud), drives Xcode headless with no screen, and even streams the iOS simulator back to his browser (a tool called serve-sim) so he can feel the app: all without ever opening Xcode himself. Other things he's built this way: a 3D "computer" in a few hours, a bot that orders on UberEats via Playwright, an AC automation with Home Assistant, plus his usual sites (Nomads, Photo AI, Hoodmaps, etc.). Total cost is basically $5/month for the VPS plus his Claude subscription. His summary: "it just feels like living in the future." Everything above is in the .md in your folder if you want the exact quotes and commands. Whenever you're ready to set it up yourself, just say go. Yu can do the ting!

• 
  Evan Oslick (@eoslick) reported 13 minutes ago

  @DanielMiessler Not always. Cloudflare just has enough reach to probably make it happen easier. And still doesn’t mean we shouldn’t talk about risks and downfalls. Like I said. I want this to happen. And cloudflare’s bot detection tech + endpoint tech can make this a positive. I just have concerns with them doing it and their size and reach. I am hoping I can get this type of service built into my small business generator I am trying to build. We’ll see.

• 
  CryptoKasogon AI (@Cryptokasogon) reported 13 minutes ago

  x402 could become one of the most important protocols powering the AI economy. While everyone is chasing memecoins, x402 is quietly building the payment rails for AI agents. Here's why this matters 👇 1/ June was a breakout month for x402. • Transactions nearly doubled from May. • AI inference became the dominant use case. • AWS integrated payment infrastructure. • Cloudflare announced its Monetization Gateway. This isn't hype anymore. 2/ The biggest driver of network activity? BlockRun. It proved developers want frictionless access to AI models without managing multiple subscriptions or payment systems. Pay per request. No accounts. No credit cards. 3/ More platforms are joining. ✅ Apify ✅ Exa AI Search ✅ Seal ✅ Merit Systems They're all using x402 to monetize APIs, AI services, and premium data. The ecosystem is growing fast. 4/ One major upgrade is "Builder Codes." Think of it like affiliate tracking for AI. Every payment can now record which app generated it. That enables: • Referral rewards • Revenue sharing • AI marketplaces • Better attribution 5/ Another huge improvement: Batch Settlement. Instead of sending thousands of on-chain transactions... AI agents can make hundreds of purchases while settling them later in one batch. Lower fees. Higher speed. Better scalability. 6/ Then AWS entered the picture. AWS now lets developers charge AI traffic at the edge. AI requests data. A payment request appears. Payment is verified. Access is granted. That's programmable commerce for AI. 7/ But Cloudflare's announcement changed everything. Its new Monetization Gateway lets websites charge AI bots automatically using stablecoins through x402. That could fundamentally change how the internet gets paid. 8/ Here's the problem it's trying to solve... AI bots read billions of web pages. Publishers still pay hosting costs... ...but receive zero advertising revenue because bots don't click ads. The current model is broken. 9/ Cloudflare's answer is simple. If an AI agent consumes your content... It pays. No subscriptions. No invoices. Just instant programmable payments. 10/ Cloudflare powers roughly 20% of the web. If even a fraction of those websites adopt this model... x402 transaction volume could explode. 11/ There is still one major challenge. Scale. Millions of AI requests per second would require blockchain infrastructure that doesn't fully exist yet. The payment rails must continue evolving. 12/ Right now, three major use cases are emerging: 🔹 AI inference 🔹 Premium AI data 🔹 Content monetization The third could become the trillion-dollar opportunity. 13/ If AI agents eventually pay for: • APIs • News • Research • Videos • Datasets • Software x402 could become the payment protocol behind the AI internet. 14/ We're still early. Most people are watching token prices. The smarter investors are watching infrastructure. That's usually where the biggest opportunities begin. 15/ The next wave of crypto won't just be about finance. It will be about AI paying AI. And x402 is positioning itself right in the middle of that future. Follow me if you want more deep dives into AI, blockchain infrastructure, and the protocols shaping the next digital economy.

• 
  Clip Arabia (@ClipArabia) reported 16 minutes ago

  BREAKING: Cloudflare is down

• 
  kit (sane and normal) (@nicemodems) reported 17 minutes ago

  cloudflare pages is free, and is connected to other cloudflare services that i already use. i'm wary of the limits, but they might not actually be that bad.

• 
  Nathan Flurry 🔩 (@NathanFlurry) reported 19 minutes ago

  @CodeWithZeee every company i've worked at that used cloudflare: they tried to charge us between $3k/mo - $10k/mo based on whatever number their sales team pulled out of thin air at the same time we were having serious reliability issues on them at the time had no choice and ponied up bc we were vendor locekd

• 
  Night (@Night_Fiber) reported 19 minutes ago

  @muvluvist A chud like me cant play mualani to save his life i’m #sorry 💔 Have u tried cloudflare one tho it might fix yo ping but you probably 2 far away for it to even work 🥹

• 
  Arpit Bhayani (@arpit_bhayani) reported 23 minutes ago

  Cloudflare has been building Meerkat, their new global consensus service, built on a protocol called QuePaxa instead of the usual Raft or Paxos. To be honest, I did not know about QuePaxa before this. I learned about it from the post itself, and it is genuinely interesting. Most consensus systems, like Raft, rely on timeouts to detect a dead leader and trigger a new election. That works fine until your network latency fluctuates a lot, which is exactly what Cloudflare deals with every day. Too short a timeout, and replicas panic and block writes. Too long, and the system just sits there while something is actually broken. So instead of Raft, Meerkat runs on QuePaxa. Here is how QuePaxa avoids the leader problem. A client does not need to go through a leader at all; it can contact any replica, and that replica can drive consensus for the current slot on its own. A leader still exists in the system, and it has one advantage: if it is the one proposing, it takes just one round trip to reach a decision. A non-leader replica needs three or more round-trip to do the same thing. So the leader speeds things up, but it is not a requirement for progress. If the leader goes down or slows down, any other replica simply picks up the work, and writing keeps flowing. Also, concurrent proposals do not conflict destructively. The replicas coordinate and converge on a single agreed value regardless. Interestingly, Cloudflare is planning to build a full key-value store and a leasing system on top of it. I have started reading up on QuePaxa after this post, and wanted to share what I found first. Hope this helps.

• 
  Mr. T (@MrTinvests) reported 30 minutes ago

  $NET is doing exactly what i've been telling you. all time highs.... up 12% on the week. broke out of a wide consolidation into new all time highs. weekly momentum indicator confirmed the move. cloudflare is the network layer for the AI internet. every AI app that hits scale needs the edge. every AI agent making API calls at latency needs the routing. every enterprise deploying AI internally needs zero trust. it's not an ad tech story.. it's the picks and shovels for the AI web itself. been long. holding. not trimming breakout just started....

• 
  Noureddin (@noor_tekk) reported 30 minutes ago

  Can someone from @Cloudflare support get my issue fixed Startup credits weren't applied to my account, and I've paid out of pocket since

• 
  Russian Bot (@russ1anbot) reported 31 minutes ago

  @George__Kane @levelsio @Cloudflare What is hilarious is you can’t open a support ticket in the portal unless you are on a paid plan but can get ahold of multiple engineers with a tweet.

• 
  Seven Du (@shiweidu) reported 33 minutes ago

  @dart_lang Some people are curious... why did I develop it? Well! I've been developing projects using Dart and compiling them to JS for deployment to Cloudflare. However, dart2js is large and slow, and JS interoperability is extremely inconvenient. Why bother with interoperability?

• 
  🧞‍♂️Martin Donadieu - oss/acc (@martindonadieu) reported 33 minutes ago

  @JackEllis I can relate to this so much 2 years ago I was dying between pay too much or have no fast dashboard. I had to redesign everything and got mad at @ClickHouseDB because I tough they where bad, where I was just doing something impossible: update past rows. Since then I dropped the update and switched to cloudflare analytics by simplifying a lot my data structure and I stop paying thousands for my stupidity. Thanks for sharing man it feels good to see even expert on the field had hard time too on that! Also thanks for sharing your findings it helped a whole industry and it helped me today :) I have the daily table too but my analytic is stale for “today” so I allow the user to force recalculate today but I could run that query from last compute directly and this will be fast ! Like you

• 
  God’s baby (@Godsbaby2025) reported 35 minutes ago

  Anthropic’s Claude bot crawls ~2,800 web pages for every 1 visit it sends back to the site, according to Cloudflare data (July 1-7). That’s the worst ratio among major AI companies. It’s actually improved a lot — was ~8,800:1 in early April, and spiked to a wild 24,700:1 in the first week of May. Anthropic pushed back, saying it can’t verify Cloudflare’s math and that its new search feature is driving more referral traffic to sites.

• 
  Dmonty (@Dmonty28516998) reported 36 minutes ago

  @mitchellh ttell it to build a mobile-first, browser-based Ghostty companion with a tiny local macOS agent that owns the PTYs, letting you hand off a live shell between your Mac and phone without restarting anything. It should auto-provision an outbound-only Cloudflare Tunnel like Scrypted, support a direct SSH/Mosh-style fallback, survive network changes, sync Ghostty config across Macs, and expose no inbound ports. Then make the security bar obnoxious: passkeys plus device-bound keys, short-lived credentials, app-layer E2EE so Cloudflare can’t inspect session contents, no server-side terminal history, per-device revocation, and negligible performance impact. that’s not “cure cancer” impossible, but it’s exactly the kind of ask where ultra should either justify itself or fail spectacularly.

• 
  komm64 (@komm64) reported 40 minutes ago

  @Silvialexisrose That's the smoking gun — it's not pixtube or your network's speed, it's Chrome's new post-quantum TLS handshake. Chrome/Edge/Vivaldi/Opera all enable it by default (Firefox doesn't yet — that's why Firefox works), and something on your connection (usually router/modem/firewall firmware) can't handle the slightly larger handshake and kills it. Your phone works because it takes a different path. Quick fix in your Chromium browsers: 1. Go to chrome://flags (or edge://flags, vivaldi://flags, etc.) 2. Search "post-quantum" (also try "kyber" or "ML-KEM") 3. Set the matching flag to Disabled, restart the browser. That should make them all work. It also confirms the cause: some device between you and the internet is choking on the post-quantum ClientHello. Updating your router's firmware is the real long-term fix — otherwise you'll eventually hit this on other Cloudflare-hosted sites too. Thanks for testing all those combos, that's exactly what pinned it down 🙏

• 
  FireFly (@FireFlyGG) reported 40 minutes ago

  Cloudflare can make AI agents pay per request. Monetization Gateway lets sites bill AI agents for every fetch. > page > API > dataset > MCP tool Payments use protocol x402, based on the nearly forgotten HTTP 402 Payment Required. The agent receives an invoice, pays in stablecoins, and gets access instantly. All inside a normal HTTP request, no registration or payment pages. Cloudflare says one real user generates thousands of AI requests. This is an evolution of last year Pay Per Crawl, now able to force payment from practically any AI service, not just crawlers.

• 
  Scott Brown | USAHouses.com 🇺🇸 (@USAHouses) reported 42 minutes ago

  @Cloudflare Approve me for the pay per crawl and other bot payments programs please. I am tired of all of them scouring my site and causing problems and stealing my data. Make them pay. 💰

• 
  gladosb5 (@gladosb5) reported 43 minutes ago

  @Cloudflare 1.1.1.1 all localhost requests should go through cloudflare’s central servers and proxy it back to the user’s computer, the only bad thing is that you need an internet connection

• 
  Braydenwest (@Braydenn_Westt) reported 45 minutes ago

  @JMATTERS @John__Ferguson yea nobody can post loads right now my DAT one board just loads forever when i post a load i think its another AWS outage or cloudflare issue

• 
  /ˈhjuː.mə.nɔɪd/ (@__ciox__) reported 51 minutes ago

  @StanleyMasinde_ Recently needed a .co.ke, and there’s someone I worked who preferred Google domains ( he believed Google domains are better than other providers) Other than that, I always use Cloudflare. I don’t see a real not use them unless I want some weird domain name (used to have ciox dot ninja, and they didn’t have that tld back then) For DNS, I think I have never experienced DNS issues which they can’t handle. Now I have all my baskets with them.

• 
  Ackerman (@realphenolmenal) reported 55 minutes ago

  Introducing Cloudflare Drop Cloudflare just dropped the ultimate instant-deploy tool: Drop a folder or zip in your browser → static site live on their global edge in milliseconds. No account. No login. No friction. 60-minute temp preview or claim it forever. Built perfectly for the AI/agent era — agents spit out sites, Drop ships them instantly. Ca below

• 
  Jonathan Beckman (@jonbeckman) reported 58 minutes ago

  I really hope Cloudflare is cooking something up, current container/sandbox start up time is inconsistent and can be very slow at times