Is Cloudflare down?
No problems detected
If you are having issues, please submit a report below.
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.
Full Outage Map
E-mailProblems 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.
-
Cloud Services (46%)
-
Domains (30%)
-
Hosting (11%)
-
Web Tools (10%)
-
E-mail (3%)
Live Outage Map
The most recent Cloudflare outage reports came from the following cities:
| City | Problem Type | Report Time |
|---|---|---|
 Aguascalientes |
Domains | |
 Köln |
Domains | |
 Trondheim |
Cloud Services | |
| Frankfurt am Main |
Cloud Services | |
 Derry |
Cloud Services | |
 Paris |
Web Tools |
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:
-
Naemt
(@monanaemt) reported
@14HWaffen88 @Blazingoth @JohnAi111 Yes it will *****, wtf are we supposed to do when cloudflare goes down for the 90th time a year and you got a massive rager.
-
Med
(@itsmedagui) reported
@jonathanzliu i can see this becoming a meme whenever there is a cloudflare outage
-
Marco Pfeiffer
(@TheTrueNemo) reported
@dominikkoch There are three problems: 1. All MCP clients are currently pretty bad. Manus likely has the best implementation. Cloudflare has a similar one, where MCP tools are exposed as code rather than direct LLM tool calls. 2. Most servers are terrible. They simply expose the API as tools without building a sort of interface for the language model, making most MCP servers very inefficient and unobvious to use. 3. The protocol itself is overly complicated because Anthropic initially tried to do too much with it and did not consider HTTP as a transport protocol in early versions. You are right, of course; it is still the only real connector interface we have, with no direct competitor yet. All of these issues are solvable and are being solved, but the hype cycle is probably over. That might actually be a good thing.
-
David
(@deivcodes) reported
@eastdakota cloudflare is going down again soon....
-
salt
(@saltjsx) reported
@fence__post *us-east-1 and i’ve never seen any serious company use anything other than aws. maybe cloudflare.
-
Polícia Federal
(@policiafoderal) reported
@dok2001 @eastdakota @Cloudflare Anycast as a service with tcp passthrough? 😌
-
Hazli Johar
(@hazlijohar) reported
@dok2001 @Cloudflare double down on email
-
John
(@johnnprofits) reported
@jonathanzliu This might be why cloudflare keeps going down 😭
-
Tridev Gurung
(@tridevgurung) reported
@eastdakota It is so over when cloudflare goes down.
-
0xMassi
(@0xMassi) reported
@johnsontrades_ @kanavtwt Bro because you never encountered a bot protection xD, try to scrape data from a website with perimeterX, Akamai or Cloudflare. Let me know then
-
Raj Agrawal
(@KyaRajKarPayega) reported
Day 0 - Dockerized Bot cloudflare tunnel with app - Config issues fixed for meta api - Went through consistent hashing This won't be just me blabbering about my day. Shitposting is ever constant of life anyways 🤧
-
Manu.ts
(@Neolectron) reported
@_smontlouis @ZeldriFR @LLCoolChris_ Cloudflare down, can't go backwards
-
David B
(@DavidB13) reported
@stage48 The forum has now been inaccessible to all users for TWO full weeks, could somebody PLEASE look into the Cloudflare issue ASAP?
-
M Pattabiraman
(@pattufreefincal) reported
Acc to Cloudflare about 25-35% is from bots and rest humans. Can never reconcile different sources. As long as there is growth ...
-
rizki
(@rizkiakiki1995) reported
SoSoValue Service Restored After Cloudflare Outage Earlier today, a global Cloudflare disruption impacted access to SoSoValue and many major websites. We’re happy to confirm that services are now fully recovered. @SoSoValueCrypto
-
markel Turku
(@MarkelTurku) reported
@Cloudflare We humans must verify whether you are actually human. We never visit sites that forced us fake cloudfare gang—we hate them now! Having trouble? Send feedback to us! That's another lie.
-
Julien
(@JulienCoulaud) reported
So often when you share an idea, someone goes "yeah but you can already do that by doing A, B, C". Sure. You can. Like yeah, you can spin up a server, configure nginx, manage infra, handle updates, break things, fix them, repeat. Or you just click deploy on Vercel / Cloudflare and focus on literally anything else. People really underestimate the power of a simple UX. And how much time saving is worth. Funny thing is, they do this in every other part of life. They buy prepared food instead of raw vegetables. They pay for convenience all the time. But somehow paying $10-20 / month for a tool that saves hours is "not worth it" because in theory you can do it yourself.
-
David Robertson
(@davidrobertson) reported
@CloudflareHelp Case #01897448 – Emergency: I am locked out of my Cloudflare account. An expired domain is causing an email bounce loop, preventing me from updating billing information. I risk losing the domain permanently. Support is unreachable from two separate emails. Need urgent escalation. DM open. @dok2001
-
Cyril Alex 🪐
(@Cyril_ALPHA) reported
3/4 Other Leaks & Scams to Watch: Fake airdrop sites draining wallets (common phishing, NOT official). Past temp outages (e.g, Cloudflare global issue fixed, funds safe). High traffic slowdowns on SSI/SoDEX launches. No major hacks/exploits on platform itself have been reported.
-
AR
(@arxdirx) reported
@EckertAnthony @anpaure @thestonechat I get it, it’s optimizations but those optimizations listed happened in 2018. By 2018 the game had sold 154 million copies. Foresight is everything. I bet you’d consider the architects at CloudFlare, Google, AWS, etc. great engineers. Yet at least once a year they discover a critical bug in their code/infrastructure that takes down the entire internet. Stuff just happens, it’s software.
-
Elijah Bowie
(@elijahbowie) reported
@techMosavvy @adocomplete @trq212 I’ve seen that, honestly have no idea what to do with it. Doesn’t even have a proper readme file. Cloudflare should create a standard base template that follows Anthropics guidelines for hosting the SDK E2B had one for Claude code CLI but never updated for Claude Agent SDK
-
zeb
(@zebassembly) reported
@tomhaerter 360 cores, which isn't actually a lot for Cloudflare but for the job this service is doing (deserializing data and serializing it in another format) it feels like a whole lot of cores.
-
Ian Nuttall
(@iannuttall) reported
@jamesvanderhaak when i say migrated and pushed, i mean gpt 5.2 built the app, set it up on a vps for me and fixed the ssl and database issues all i did was change the ip address in cloudflare
-
Harmis Tervadiya
(@HarmisTervadiya) reported
I think I don't know how to pitch I pitched 3 project ideas to my faculty for college project, he said this try to find something more. Ideas: - Real Time Local Problems sharing site(Community Base) - Learnyst clone - A service platform like Cloudflare for hyped product drops Going again today with a more simpler idea, let's see how it goes
-
ViktiniPlay 🇷🇺♥️🇺🇦
(@ViktiniPlay) reported
It has never been said that there will be fair competition. As a result, we got a lot of blockages. The Cloudflare service, which runs almost all websites in the world, was also blocked. It's just terrible.
-
Raju Khunt
(@raju_khunt) reported
Took a vacation.. Cloudflare went down.. Even when I rest.. the internet reminds me why redundancy matters. Always have a backup.. for your backup.. for your backup. #SaaS #developer
-
Emre Baş
(@EmreBasTweets) reported
@dok2001 @Cloudflare email service.
-
The Deal Director
(@thedealdirector) reported
@eastdakota This is the new copy pasta meme whenever half the internet goes out due to a Cloudflare outage.
-
Konstantinos
(@kostasbotonakis) reported
@dok2001 @Cloudflare 1) Proper email service, not just filtering / routing incoming. 2) Other DB solution based on PostgreSQL 3) OAuth that can be used outside of Zero Trust , like Clerk or other similar services. 4) AI Gateway but BYOK and not just OpenAI.
-
developing valhalla - h/acc
(@valhalla_dev) reported
Today I'm testing something fairly complex with Claude Code to see how it debugs things. Interestingly, preliminarily it's not great. It's debugging two Rust projects and a CloudFlare Worker at the same time, and it is missing some fairly easy things (capitalization of JSON fields, for example) and has already run long enough to have to compact twice. I saw the issue a while ago, and even before that I ran through a much more logical progression on how to fix it than CCode has. This makes sense to me, because one of the things that comes with seniority in SWE is being able to logically run through debugging processes. It comes with experiencing lots of very specific bugs in processes and code. This means that one of the things that can make you safe and more powerful as a dev with access and knowledge in using AI tools is to know when in a debugging process to turn off the AI and take the wheel. What did it do well? - connected to infra, added an nginx route, changed firewall rules, installed and migrated a postgres DB (basic DevOps) - created very basic API endpoints in Axum (Rust) and in Node (CloudFlare Worker) and deployed both (basic DevOps + junior dev code) This is the first instance I can honestly say I could have probably done 80% of this faster, but up until the bugs that held up the debugging process for a while, it was moving faster than I could, AND I was able to get other work done while it ran.