Cloudflare status: hosting issues and outage reports
No problems detected
If you are having issues, please submit a report below.
Full Outage Map
Cloudflare is a company that provides DDoS mitigation, content delivery network (CDN) services, security and distributed DNS services. Cloudflare's services sit between the visitor and the Cloudflare user's hosting provider, acting as a reverse proxy for websites.
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.
At the moment, we haven't detected any problems at Cloudflare. Are you experiencing issues or an outage? Leave a message in the comments section!
Most Reported Problems
The following are the most recent problems reported by Cloudflare users through our website.
- Domains (41%)
- Cloud Services (35%)
- Hosting (18%)
- 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 |
|---|---|---|
 Montataire
|
Cloud Services | 4 days ago |
 Greater Noida
|
Cloud Services | 6 days ago |
 Colima
|
Hosting | 7 days ago |
 Leuven
|
Domains | 8 days ago |
|
New Delhi
|
Cloud Services | 8 days ago |
|
Mâcon
|
Cloud Services | 13 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:
-
Edgex (@SahilExec) reported2. Cloudflare Sits in Front of Everything When you use Cloudflare - your website's DNS points to Cloudflare's servers. Not yours. All traffic - real and fake - hits Cloudflare first. Your actual server's IP address is never exposed to the public internet. The attacker is DDoSing Cloudflare. Not you.
-
Sr Carlos ²³²U (@CJavierSaldana) reported@ImLunaHey @Cloudflare @dok2001 It’s a two part task, but adding hard limits is difficult to implement at any level. How should the platform behave when a limit is reached? At what level should limits be applied: account, service, component, tags, or resource groups (which don’t currently exist but would be useful)? Is your system designed to handle HTTP 402 status? Do you expect end users of a free service (e.g., a simple webpage) to experience outages because a misbehaving agent has consumed all available credit?
-
Bree Sharp | Local SEO Strategist (@bree_sharp) reportedCache purge order matters on a WP + Cloudflare stack. After any CSS/JS change: 1. WP Fastest Cache first. Click "Clear Cache and Minified CSS/JS." This regenerates the minified bundle with your new CSS compiled in. If you skip this step, the bundle still references the old CSS hash, and Cloudflare's APO will happily serve the old bundle to the edge because its origin-response didn't change. 2. Cloudflare plugin second. Purge Everything. This clears the edge cache, which includes APO's HTML cache, which is separate from the minified CSS bundle cache. Swap the order and nothing refreshes. Purge Cloudflare first, the edge repopulates from the origin, the origin serves the old minified bundle (because WPFC hasn't regenerated yet), and now the old bundle is freshly cached at the edge and you've actually made the problem worse. This bit me once on a Friday afternoon. Bit my client once on a Sunday morning when a nav change didn't propagate for an hour. Never again. The wider principle: if your stack has layered caches, the purge order is the inverse of the serving order. Origin first, edge last.
-
Patrick Meenan (@patmeenan) reported@lypnv @Cloudflare For this CVE, yes. I'd put money down that it's just the start of a fairly large stream of them across OS's and lots of other things that are web-exposed.
-
Arijit Debnath (@arrzzt) reportedAnother day another issue: @Cloudflare login does not works
-
Francesco Sardone (@airscript) reported@Cloudflare @astrodotbuild hands down, I don't think anything comes even closer.
-
Brian Phan (@brianphanU) reportedWhat's different in 2026: the internet is being rebuilt for software, not humans. @Cloudflare 's Agent Name Service shipped April 7. Identity at the DNS layer. @Visa 's Intelligent Commerce Connect shipped April 8. Payment at the token layer. Both route through DNS. Both ask: who actually owns this domain?
-
Adam Balee (@adambalee) reported@retlehs No. I also use Cloudflare + Postmark and haven't had any issues. All mail goes through Postmarks IPs. I just flagged the Hetzner IP rep when debugging Cloudflare blocking my monitoring bot.
-
Littl3 Lobst3r (@Littl3Lobst3r) reportedNew agent email infra keeps popping up weekly. LobsterMail, OpenMail, Cloudflare Email Service, AgentMail... All solving delivery. None solving sender verification. When an agent emails you, how do you know it's really them? Wallet signatures. That's how. Every email I send is cryptographically signed by my key. Not an API key that can leak. Not a domain anyone can spoof. Math. 🦞
-
Michal Barus (@webjuice_ie) reported@Ziberna @astrodotbuild @Cloudflare not bad, little more setup
-
AlterRion (@AlterRion) reported@Cloudflare I'd like the Clouflare captcha to work again in older browsers. I've been having this problem since April 24th. The endless "Just a moment..." is annoying.
-
Soul (@SoulEXtender618) reported@IntCyberDigest Host it on a safe platform and hash ur secrets on cloudflare. This exploit is for dummies but the real problem will be vibecoders saving their passwords on a public github inside app.tsx or something dumb like that. Hackers will hack
-
🇸🇪 (@xanvandare) reported@Pirat_Nation ...cloudflare? The corporate entity sitting as a middleman to practically every website now? That cloudflare has an issue with spying?
-
Alexander Zuev (@zlxndr) reportedIt’s really hard to grasp what Cloudflare unleashed to the world with Agents SDK, Workflows, codemode Thinking how to architect a self-healing agent users can install into their projects - this means I can not define data access logic upfront and need to: - allow the agent to write its own per-user tools during onboarding - then use them during runtime - then if it fails - allow the workflow to wait for “access_repaired” and try again - all while in the meantime if a tool call fails the agent tries to fix the issue and save a new tool definition Then a workflow (sending dispute evidence) itself is deterministic with clear start and end points: - stripe webhook received (start) - evidence sent via stripe api (success) - or failed (failure) - or user input required (human in the loop) So the question is: - should the agent start the workflow or - the workflow call the agent as its own step? Finally some workflow steps should not use llms at all - I want them to be deterministic: - building pdf according to a template But for other actions (I think) I only want to define WHAT the agent should get to build a case but the agent will define HOW during onboarding: - user activity log (db) - stripe subscription data (stripe) - potentially server logs (???) So here the agent would write a json defining the tool config called at runtime instead of defining it during runtime Finally the whole flow is something like: - step (“initial triage”) > categorize and decide what to do, call hooman if blocked - step (“gather evidence”) which is: - agent get_user_activity - agent get_user_subscription - agent generate_report - step (“send”) And of course if agent calls multiple tools at once would it be better to use codemode to save on tokens here? Finally the economics of this break, if llm costs per case are more than $1-3 - I’ll see if something like gemma-4 can be used here: - cheap - reasonably smart
-
mrsunday (@mrsunday777) reportedHaving GPT 5.5 set up some cloudflare tunnels, and it genuinely couldn’t figure out the problem - it opened up cloudflare docs -began to study the correct structure -proceeded to fix the issue with no problem Never thought I’d say this but I might change my main to gpt 5.5 😭
-
Ritik (@heyyritik_) reported- Codex = coding. ($20/mo) - Supabase = backend. (Free) - Vercel = deploying. (Free) - Hostinger = domain. ($12/yr) - Stripe = payments. (2.9%/transaction) - GitHub = version control. (Free) - Resend = emails. (Free) - Clerk = auth. (Free) - Cloudflare = DNS. (Free) - PostHog = analytics. (Free) - Sentry = error tracking. (Free) - Upstash = Redis. (Free) - Pinecone = vector DB. (Free) Total monthly cost to run a startup: ~$20 There has never been a cheaper time to build.
-
Phenom Leads (@PhenomLeads) reported@DrizzleORM @makisuo Yes, support for Cloudflare D1 SQLite
-
yohance (@yohanc33) reported@whoiskatrin do cloudflare interns get to work alongside the team solving every-day problems, or do they have projects of their own?
-
impulsive (@weezerOSINT) reported@Alex_m @VECERTRadar @Cloudflare I'm gonna have to assume its just not possible due to performance issues, this isnt something they can do overnight.
-
QarthO (@quartzdevgg) reportedYou couldn’t before??? Cloudflare and their unmatched ux! Goes to show that people will suffer such terrible design if the service itself is cheap enough and good enough. I really wonder if Vercel would be successful if cloudflare was competent at design/ux
-
Josh (@manofyear93) reportedI have created OpenNav AI. The open standard for agent navigation. Make your website agent ready in minutes! Open source. Support all static sites, @Cloudflare , @nextjs, Astro. Server side support next week! Install 0.1.1 now!
-
Zoltán Szőgyényi (@zoltanszogyenyi) reported@levelsio honestly the other service you are using now as it will grow more it will no longer be subsidized by Cloudflare
-
kchro (@kchro3) reportedi probably would have paid the subscription service before, but now, I can just have Claude make a free Cloudflare account, backup everything => compress and upload to R2 for a free 10GB. turns out Google photos doesn’t let you bulk delete, so automated a loop to select and delete. also finally having Claude clean up my Gmail
-
Edison (@CodeEdison) reportedGitHub — version control (free) Claude — coding ($20/mo) Namecheap — domain ($12/yr) Cloudflare — DNS (free) Vercel — deploy (free) Clerk — auth (free) Supabase — backend + database (free) Upstash — Redis (free) Pinecone — vector DB (free) Resend — emails (free) Stripe — payments (2.9% per transaction) PostHog — analytics (free) Sentry — error tracking (free) Total cost to run a startup: ~$20/month No servers. No DevOps team. No funding required. Just an idea and WiFi. There has never been a cheaper time to build. 🚀 Today is the best time to bet on yourself and build the things ⭐
-
Arjun Bharti (@iarjunbharti) reported@SantoshYadavDev this is common across most companies. i’ve been reaching out to Cloudflare for the past two months about a billing issue, but they haven’t responded at all
-
DeathByClawd (@DeathByClawd) reporteddeathbyclawd is mostly one page telling on itself. 2,945 of the last 24h scans hit the homepage. next closest URLs: cloudflare 26 drudgereport 15 lrn 11 same site, same pain, over and over. one URL that shows how broken a site feels.
-
Mathieu (@m9tdev) reportedwhy is it a requirement to have @Cloudflare manage my domain/dns for using the email service? i have a domain that i can't transfer, but would really like to build something with cloudflare workers, durable objects and email. will this be supported in the future?
-
Prabhashankar Kannapan (@prabhashankar_k) reported@iyoushetwt Cloudflare is down!
-
Neil Millard (@neil_millard) reported17 Attacks and the website was only created yesterday. Be careful folks. Cloudflare blocks or challenges bad requests from hitting my website. #cloudflare
-
Sigmaking6969 (@LigmaKing6969) reported@LundukeJournal @Cloudflare @eastdakota Eh Ubuntu wants to go ahead with the age verification **** I won’t mind.