1. Home
  2. โฏ
  3. Companies
  4. โฏ
  5. Cloudflare
Cloudflare

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.

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

Live Outage Map

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

CityProblem TypeReport Time
Paris Cloud Services 2 days ago
New York City Hosting 4 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:

  • PedroGuiti
    Pedro Guitian (@PedroGuiti) reported

    if you're building a startup. pause for a second. You should stop overpaying for your stack. this is enough to launch: claude - coding supabase - backend vercel - deploys GoDaddy - domain stripe - payments github - version control resend - emails clerk - auth cloudflare - dns posthog - analytics sentry - errors upstash - redis most of this is free. The real cost is time, so ship fast, and optimize later

  • YourPope2026
    Your Pope (@YourPope2026) reported

    @DanNeidle Ask Cloudflare if they can help?

  • SyAzCrypto1
    SyAz (@SyAzCrypto1) reported

    $12M fees in 12 days, then gone. @Noxa_Fi Robinhood's launchpad that ran entire meme season just switched itself off & the whole ecosystem repriced 30% overnight. I posted the $85 โ†’ $1.79M cash-cat:native run a week ago. And here's how the story ended, in numbers. The Rise, 12 days: โ†’ Robinhood Chain launches july 1, built for tokenized stocks and RWAs โ†’ Noxa becomes the launch engine, 60,000 tokens, 75% of all deployments โ†’ $3.1B first-week DEX volume, 200k active addresses โ†’ Out-earns @Pumpfun five days straight, 4x at one point โ†’ cash-cat:native peaks near $226M, doing $98M daily volume, 17% of the whole chain The Exit, 5 days: โ†’ july 11, CASHCAT hits peak volume, Noxa pauses all launches the same day โ†’ july 13, website goes dark, blamed on a Cloudflare issue โ†’ july 15, team quits collecting fees, hands revenue to creators โ†’ cash-cat:native drops 33% in 24hrs, wicks $0.19 to $0.08 โ†’ $DIH, $HOODIE, $HOODRAT all down 30%+, new deployments collapse The strange part, nothing was stolen: โ†’ Liquidity locked forever on Uniswap, even Noxa couldn't touch it โ†’ Everyone got their fees & no funds taken So nobody stole anything, but retail still ate a 30-60% draw-down. The timing did what a rug does without being one. What this actually teaches: โ€ข One website was the whole ecosystem. The site went down for two days and a $226M token lost a third of its value โ€ข Locked liquidity protects you from theft, not from attention leaving โ€ข The RWAs Robinhood built the chain for sits at $12.7M total, cash-cat:native alone peaked at 12x that. So yeah the speculation gets you users but it never keeps them. The launchpad proved it in 12 days flat.

  • WVROfficial
    W V R ๐Ÿ‘Š๐Ÿผ๐Ÿฆพ (@WVROfficial) reported

    I had to make some changes today and it costed me a MONTH of Codex Usage! I think itโ€™s worth it for me to talk about it - make sure this doesnโ€™t become you!!! So, Iโ€™m a startup founder, just like all of you guys!! We do websites as one of our many services - like a lot of the people here on TPOT where tech lives. Why wouldnโ€™t we? Itโ€™s easy, we can outclass competition on speed, much more. Anyway - thatโ€™s not the point. The point is that i serve my sites with a wrapper that gets served over my host. That host uses ESBuild. To ship stuff that wonโ€™t compile on ES Build, and on occasion just for like more complex websites like 3D sites or heavy SEO sites with a lot of files and assets - i use a CDN. Works great, totally fine. But I realized today I had a client siteโ€™s files stored on an R2 bucket that was on the clientโ€™s domain. In this case the big issue with that is my own IP! We make the content for them, do their SEO, their communications, and more - and i very stupid it was serving everything from a CDN that was on domains I donโ€™t own and control via my Cloudflare. In human terms that means my client could say โ€œ**** youโ€ tomorrow and walk away with the extremely robust SEO machine I built them. So I had to spend almost a half a month worth of codex credits today to fix it ASAP. All I can say is that I wonโ€™t make that mistake again - even though it never hurt my business - it could have! And that matters.

  • KhafraDev
    Khafra (@KhafraDev) reported

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

  • TheAnirudh
    Anirudh Coontoor (@TheAnirudh) reported

    @jullerino Sometimes it keeps refreshing the page, happened multiple times with the cloudflare dashboard login page. I wanted me to login but it kept refreshing.

  • Porkbun
    Porkbun (@Porkbun) reported

    @liltechnomancer @joshmanders Was the subject of the support email: "I put the cloudflare namservers in before my domain expired and it never transferred."? If so, there may be some confusion about how domain transfers work. You don't transfer a domain by updating it's name servers, you have to initiate a domain transfer at the gaining registrar using the domain's auth code. There is no order block on your account so it isn't remotely an issue like Josh experienced two years ago. Your domain is simply expired and needs to be renewed or transferred. Since it's almost 30 days expired, you'll want to act quickly. If that's not the correct ticket let me know.

  • katewerk
    Katewerk (@katewerk) reported

    @Cloudflare Adding @Cloudfare broke my site so badly I had to suspend it within days. Now, despite cancelling my paid subscription, and removing the domain, your website won't allow me to remove my credit card from your records -- and your support bot is ghosting my tickets. Fix it.

  • UK_Daniel_Card
    mRr3b00t (@UK_Daniel_Card) reported

    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

  • mohitdotdev
    Mohit (@mohitdotdev) reported

    Encrypt docs & PII end-to-end on Cloudflare Workers: client wraps random DEK with your password key, server adds outer master-key wrap; share securely by re-wrapping the DEK for others, grant edit rights with consent - server never sees plaintext. This is my way forward to keep every bit of user data encrypted. Fully and provable.

  • e_tartakovsky
    Eugene Tartakovsky (@e_tartakovsky) reported

    Googlebot still reads the web about as much as every AI crawler combined. Per Cloudflare, in July 2025 Googlebot was 39% of crawler traffic, while GPTBot was 12%, ClaudeBot about 10%, and Meta's bot under 8%. Measured across a full year, Googlebot made up 4.5% of all page requests and every AI bot together 4.2%. One crawler roughly equals the whole field. The raw fetch counts say the same. On Vercel's network in one month, Googlebot made 4.5 billion fetches, against 569 million from GPTBot and 370 million from ClaudeBot. Demand is smaller than the noise suggests too. Per Pew, 34% of US adults have ever used ChatGPT. Most searching still happens the old way. This is worth saying because companies are now writing files and changing configs specifically for AI bots, sometimes blocking the crawlers that send them the most readers. And Googlebot is the only one of these crawlers that renders JavaScript, so it is also the strictest reader to satisfy. The work that makes a page readable to AI is the same work that has always made it readable to Google: finished HTML from the server, a clean sitemap, content that does not hide behind scripts. There is no separate AI project waiting to be funded. There is the foundation you already owed Google. And Google still does most of the reading.

  • marsuplamy
    Marsu (@marsuplamy) reported

    The Agentic Economy 2024 was the year of LLMs. ChatGPT, Claude, Gemini and models like them responded to prompts and that was enough. But 2025 brought something different. Models were no longer just responding, they were planning, calling tools, executing code, coordinating with other systems, and doing all of this without constant human oversight. This transition transformed AI from something you query into something you delegate to. This is called agentic AI and with it an entirely new economy began to take shape. To understand the scale of this economy a few numbers are worth looking at. The AI agents market is expected to grow from 7.84 billion dollars in 2025 to 52.62 billion dollars by 2030, at a compound annual growth rate of 46.3%. McKinsey projects that agentic commerce could orchestrate between 3 and 5 trillion dollars in global revenue by 2030. These numbers are not theoretical, the infrastructure is already being built. In just the six month period between April and September 2025, Visa, Mastercard, PayPal, Stripe and Google all launched agentic payment infrastructures. So what are these agents actually doing? They are purchasing services on your behalf, paying other agents, accessing APIs, buying data, and doing all of this while making decisions in fractions of a second. Stablecoin transaction volume reached 33 trillion dollars in 2025, up 72% year over year, with supply surpassing 300 billion dollars. Agentic payments and machine to machine payment flows are cited as one of the key drivers behind this growth. Stablecoins like USDT and USDC are critical for agents because they allow programmable payments without price volatility. But legacy payment infrastructure was never designed for this world. Credit cards require human authentication, subscriptions demand upfront commitments, and API keys depend on manual onboarding processes. All of these systems were built for humans. When millions of agents are making countless payments per second none of these systems work technically or economically. x402 and the Awakening of HTTP 402 When web standards were being written in 1991 HTTP status code 402 was added and defined as 'Payment Required'. That day it was reserved, set aside for future use. This code waited more than thirty years and when its future arrived it turned out not to be human. The x402 standard activated this dormant code as a native payment layer for the internet: a server responds to a request with 402 and a price, the client pays on-chain in stablecoins, retries the request with proof of payment and receives the service. No account creation, no card on file, no subscription, no human. The protocol was launched in September 2025 by Coinbase and Cloudflare through the x402 Foundation. The coalition behind it is unusually broad, Google, Visa, AWS, Circle, Anthropic and Vercel are all core foundation members. Within five months of launch x402 had processed over 100 million transactions. In a single week in October 2025 the protocol handled approximately 500,000 payments, a 10,780% increase from the prior month. The technical side of x402 is very clean. For a developer integration is a single line of middleware, set a price per endpoint, point to a facilitator, and the API can charge per request in stablecoins. When an agent wants to access a service payment happens automatically inside the HTTP request, settlement completes within the round-trip. Zero human intervention. The Problem: Public Rails Don't Work for the Agentic Economy Now we come to the critical question. When millions of agents make transactions and every one of those transactions is visible on a public blockchain, what happens? Which APIs the agent uses, which data it accesses, which services it purchases, how much it pays, who it works with, all of it becomes completely visible. This is not just a user privacy problem, it means the strategy and logic the agent operates on is open to competitors. Is a company's agent feeding from the same data source as a rival's agent? How much is it spending on which compute services? How are supply chain decisions being made? All of this becomes readable on a public chain. On top of that there is the gas fee problem. On Ethereum and Tron fees shift constantly with network congestion. If an agent is making hundreds of microtransactions per second modeling your unit economics becomes impossible because you have no idea what costs will be in advance. For the agentic economy to work payment rails must be both private and predictably priced. Why Bitcoin is the Neutral Rail There are several clear answers to why Bitcoin stands out as the ideal settlement layer in this equation. First, censorship resistance. No central actor can stop, censor or restrict agent payments. For agents to operate autonomously the payment infrastructure must also be autonomous and uncensorable. Second, deterministic finality. Bitcoin's proof-of-work security is the most battle-tested and proven consensus mechanism in existence. For agent payments settlement must be definitive and irreversible. Third, global liquidity. Bitcoin is accessible everywhere in the world with no geographic restrictions and agents operate without borders. Fourth, the UTXO model. Unlike Ethereum's account-based model Bitcoin's UTXO structure allows non-conflicting transactions to be validated in parallel, a natural advantage for high-frequency agent payments. Where @Utexocom Fits The layer that combines Bitcoin's advantages with USDT and makes it production-ready for the agentic economy is Utexo. The RGB protocol issues and transfers USDT as a native asset on Bitcoin's own layer. Transfer details never get written to a public ledger thanks to client-side validation, only cryptographic commitments are anchored to Bitcoin UTXOs. So when an agent makes a payment who sent what to whom never leaks outward. The Lightning Network allows these assets to settle in milliseconds, at around 200ms latency. Utexo handles channel management, liquidity and routing entirely internally, with fees fixed and predefined at the protocol level. For the agentic economy this combination means the following. The agent pays in sub-second time, costs are predictable, payment details are private, and Bitcoin's finality provides the settlement guarantee. With the Mint component USDT from Ethereum, Tron or Solana can be moved onto Bitcoin rails. With the Swap component non-custodial exchange between BTC and USDT is possible. And the SDK reduces all of this complexity to a single API call, meaning a developer integrating agent payments never has to run a Lightning node or manage RGB infrastructure. Tether not only supporting this infrastructure but leading the seed round themselves, and preparing to issue USDT natively on Bitcoin through RGB protocol v0.11.1, answers the question of which rails the agentic economy will be built on. Machines are making payments now. Those payments need to be private, predictably priced, and anchored to Bitcoin. The infrastructure is here.

  • YourPope2026
    Your Pope (@YourPope2026) reported

    @CyberSecAJ @DanNeidle Agree it's nothing to do with Cloudflare. It's a community project. But Cloudflare will have experience of their customers being maliciously added. So they might be able to help.

  • onlyawassie
    OnlyAWassie (@onlyawassie) reported

    @CryptoTomYT @vladtenev Lockin? builders getting Problems because of the copyrights from Rh cloudflare shutting them off if they get reported how you wanna Build something on this chain

  • dezign_ash
    Ash Designs (@dezign_ash) reported

    @BraedendotTECH We're not as sharp as we were "before AI". We're just one @Cloudflare outage away from realizing how dependent we are on AI.

  • keith_c3529
    Keith C Wenzel (@keith_c3529) reported

    Never be desperate enough to buy tickets through @StubHub or @axs โ€ฆ.one of the worst customer experiences ever. You would be incorrect to think all those extra fees have gone into a good customer experience or a mobile application that actually works. Their inability to staff and manage their @Cloudflare leaves them inept at servicing customer access to their paid for tickets. Instead they force you to jump through hoops, stand in lines and waste your time at an event since they refuse to fix their stuff or hire competent employees who know how to properly configure their cloudflare. Pathetic.

  • Porkbun
    Porkbun (@Porkbun) reported

    @liltechnomancer @joshmanders That looks like something Cloudflare requires, not us. They appear to have a transfer process that includes extra steps because they require the domain to be using their DNS. It looks like the actual transfer was never submitted to Cloudflare because they were / are waiting on this step. If you changed your name servers after expiration, 10 days after expiration we disrupt DNS as required by ICANN and that would have prevented what Cloudflare seems to be waiting for. It's honestly best to renew / transfer before expiration to prevent wonkiness like this.

  • GuruVerseX
    GuruVerseX (@GuruVerseX) reported

    @ProMint_X @Noxa_Fi Cloudflare would help lol

  • enginoid
    Fred Jonsson (@enginoid) reported

    I went with serverless SaaS for small-scale company infra and building out products, and I basically regret it. The infra I have is fairly light. They are mostly key primitives I use in my contracting work and research, such as: - agents API/UI with jobs+observability-caching - container APIs for managing tier-3 GPUs - easy hosting and perf tuning of arbitrary OSS LLMs On top of that I've got a product in progress around evaluation, annotation and optimization of AI tasks, and relies heavily on those APIs. This product is for fairly sophisticated AI product companies and won't experience large-scale usage. Overall, it's quite a simple setup of maybe 10 Cloudflare Workers and Firebase for collaborative editing. My reasoning for going serverless was to minimize the gap from experimentation to shipping: pay very little monthly and just pay when it scales out. The other one was to treat the first product as an investment in building lots of other products without setup costs โ€“ get IaC, auth, authz, etc. working once and reuse it in other products. Yet another consideration was to use the wonderful high-level serverless primitives like Cloudflare Workers, Workflows and Firebase. You get a lot of good stuff that is genuinely hard to distribute, and it just works, so you can focus on product. But I think serverless was the wrong choice for me in this trade-off, for two reasons: 1. Cost predictability. Every few weeks I get a random runaway $50 bill from one service or another, maybe because the agents went a bit wild with E2E testing. I'm sure "random load spikes" have cost me as much as $300 at this point over a couple of months. This is not a lot of money in the grand scheme, but for an application with zero users, it's a lot of variable cost. This week I saw a notice from Cloudflare showing the way they'll charge for Workflows makes them uneconomical. And each time it's a bit disruptive to figure out what is causing the traffic -- all while I have zero users except myself. I still haven't got a horror story, thankfully, but I am worried that I'm due any day. 2. Phoenix environments. The other issue is that the best iteration speed I've found is when agents can 100% run the infra and replicate the full user experience, fast, locally and in memory. This is incredibly easy with postgres and a monolith, but it's messy to get reliable when you use services like Firebase and Durable Objects that aren't fully emulatable and are a bit messy architecturally to swap out. My preview deployments are really sprawled on CloudFlare because business logic is split between workers and DOs don't work well with preview deployments, so need to deploy a new set of services. So in that sense, using something high-level is far from simplifying compared to running of a k8s namespace or a single binary. I'm probably a little too used to startups and seeing what kind of issues you run into that I wanted to pre-empt with "proper infra." The indie hacker advice is to get a $5 VPS and set it up on postgres, validate and then think about proper deployment. Having shipped mostly software at scale and knowing what's needed, I still struggle with doing things that don't scale, but this advice seems right for day 0 and I should have heeded it. All this is to say โ€“ PaaS vs. IaaS are two very different directions worth taking seriously at the outset of a new project. The narrative is that PaaS will simplify things, but this hasn't been my experience so far. If anything, it has created a lot of novel constraints that complicate the architecture. And I think by outsourcing what I'd need to understand if I rolled it myself, I sold away too much freedom. So in retrospect, I should have gone with the other option I was considering: fixed resources via something like k8s + PlanetScale. If I were making the decision again, I would have upweight the cost predictability a bit more โ€“ at some level it was a consideration, but I thought I could tame it with lots of limits and budget notifications. But anytime you add a lot of mechanism (quotas and alerts) to replace an invariants (I can't pay for more computers than I have), it's probably worth paying closer attention. When I think about how difficult things like workflows, queues and even realtime are to implement on postgres and have an entire WAL to drive the whole application: I'm probably missing something, but the answer seems to be... not very hard. Maybe at millions of users, but not at hundreds or even thousands. So I'm going to try swinging the pendulum back in the direction of more control. I'm optimistic that if I don't operate the database layer itself, this will turn out to be a better foundation. And if I'm so lucky, I can one day write the post about moving off postgres. Let's see what happens after the honeymoon period!

  • bcs_erictaylor
    Eric Taylor (@bcs_erictaylor) reported

    Date: 7/15/2026 Here is an update on our updated information on our proactive efforts to heavily impact the Kali365 platform. As expected, the user "Octopus King" is working hard to rebuild is infrastructure, but we have the fingerprinting locked in and will continue our efforts. At the time of this blog posting, we have submitted an IC3 report but honestly not expecting much. I believe it will be up to companies like ours to keep submitting takedown requests as fast as we see them to limit its operations. Below is the information 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 Pending 'Takedown Requests' submitted to BL Network: IP Address: 162.33.178[.]105 IP Address: 72.5.43[.]195 Pending 'Sinkhole Requests' submitted to @Cloudflare FQDN: updateteampanel[.]xyz FQDN: privatetoken[.]app FQDN: servoquil[.]org FQDN: vuredonte[.]org FQDN: yalmorind[.]org FQDN: ondrevail[.]org FQDN: caldivore[.]org FQDN: mvpaffiliatecz[.]site FQDN: vuredonte[.]org #Kali365 #CTI #threatintel

  • ycjgt
    Ali (@ycjgt) reported

    Cloudflare DO and D1 are down down bad

  • unclebigbay143
    U N C L E BIGBAY โœจ (@unclebigbay143) reported

    Today's Engineering Concept: '๐—ฅ๐—ฎ๐˜๐—ฒ ๐—Ÿ๐—ถ๐—บ๐—ถ๐˜๐—ถ๐—ป๐—ด' ๐—ช๐—ต๐—ฎ๐˜ ๐—ถ๐˜€ ๐—ฅ๐—ฎ๐˜๐—ฒ ๐—Ÿ๐—ถ๐—บ๐—ถ๐˜๐—ถ๐—ป๐—ด? Rate limiting is the practice of restricting how many requests a user or system can make within a specific period. ๐—ช๐—ต๐˜† ๐—ฑ๐—ผ๐—ฒ๐˜€ ๐—ถ๐˜ ๐—บ๐—ฎ๐˜๐˜๐—ฒ๐—ฟ? Without rate limiting, a single user or malicious bot could overwhelm your server, degrade performance, or abuse your APIs. ๐—ฅ๐—ฒ๐—ฎ๐—น-๐˜„๐—ผ๐—ฟ๐—น๐—ฑ ๐—ฒ๐˜…๐—ฎ๐—บ๐—ฝ๐—น๐—ฒ Imagine a login endpoint with no rate limit. An attacker could attempt thousands of password combinations every minute. A simple rate limit can significantly reduce the effectiveness of brute-force attacks. ๐—›๐—ผ๐˜„ ๐—ถ๐˜€ ๐—ถ๐˜ ๐—ถ๐—บ๐—ฝ๐—น๐—ฒ๐—บ๐—ฒ๐—ป๐˜๐—ฒ๐—ฑ? Most systems track requests by IP address, user account, or API key. Once a predefined limit is reached, the server temporarily rejects additional requests, often with an HTTP 429 (Too Many Requests) response. ๐—ช๐—ต๐—ฒ๐—ฟ๐—ฒ ๐—ถ๐˜€ ๐—ถ๐˜ ๐˜‚๐˜€๐—ฒ๐—ฑ? โ€ข ๐—š๐—ถ๐˜๐—›๐˜‚๐—ฏ: GitHub's REST API limits how many requests you can make per hour to prevent abuse and ensure fair usage for everyone. โ€ข ๐—ฆ๐˜๐—ฟ๐—ถ๐—ฝ๐—ฒ: Every payment request can include an Idempotency-Key, ensuring a customer isn't charged twice if the same payment request is retried. โ€ข ๐—ข๐—ฝ๐—ฒ๐—ป๐—”๐—œ: The API enforces rate limits on requests and tokens per minute, helping maintain reliability and preventing a single application from overwhelming the service. โ€ข ๐—ซ (๐—ณ๐—ผ๐—ฟ๐—บ๐—ฒ๐—ฟ๐—น๐˜† ๐—ง๐˜„๐—ถ๐˜๐˜๐—ฒ๐—ฟ): X limits actions such as following many accounts, liking posts, posting, or sending DMs within a short period to reduce spam and bot activity. โ€ข ๐—–๐—น๐—ผ๐˜‚๐—ฑ๐—ณ๐—น๐—ฎ๐—ฟ๐—ฒ: Cloudflare lets website owners configure rules like "block or challenge any IP that makes more than 100 requests in a minute" to protect against abuse and DDoS attacks. ...and almost every public API uses rate limiting to protect its infrastructure, ensure fair usage, and maintain service availability. ๐—ง๐—ต๐—ฒ ๐˜๐—ฎ๐—ธ๐—ฒ๐—ฎ๐˜„๐—ฎ๐˜† A reliable system doesn't just answer requests. It also knows when to say "not now. It's too many from YOU."

  • PinzariAndrej
    Pinzari Andrei (@PinzariAndrej) reported

    @Cloudflare Making the exception machine-readable is a meaningful improvement over silently weakening validation. Will Cloudflare publish telemetry on how often EDE 33 is returned and how long exceptions remain active? That would help quantify both operational value and the risk of temporary bypasses becoming sticky.

  • Priya_Upadhyay_
    priya upadhyay (@Priya_Upadhyay_) reported

    @avijeet_writes changing your dns settings to google or cloudflare on your wifi router usually fixes that weird routing issue immediately

  • Ice_Cre4m_Art
    Ice_Cre4m_Arteeeee (@Ice_Cre4m_Art) reported

    @V1kov_ I assume the problem is with the Internet service provider/router but not sure I have problems with ds too (vc didnt work and other) on my wifi at home but i use cloudflare WARP or any other VPN and that kinda help

  • icyphox
    Anirudh Oppiliappan (@icyphox) reported

    @threepointone @AKuederle Not sure where all this anger is coming from. The folks on Bluesky (and me here) were poking fun at the banner textโ€”which was apparently *meant* to be tongue in cheek! Wasnโ€™t clear. Nobodyโ€™s trying to cancel Cloudflare. Letโ€™s all chill out for a bit.

  • atharva_again
    Atharva Verma (@atharva_again) reported

    Gpt 5.6 sol just deleted the *** worktree it was working in and tried reading the code from another worktree. On top of that I'm seeing frequent websocket and cloudflare errors. I saw fewer websocket errors in 5.5 and NO cloudflare errors.

  • ohfarfoxache
    Become Ungovernable ๐ŸฆŠ (@ohfarfoxache) reported

    This is the level of corruption in the Australian Government. A Comcare employee with hurt feelings, can ask the eSafety Commissioner to try and get a blog taken down because it hurts the feelings of a Comcare employee. This is what actual corruption looks like, two government pals abusing their government positions for their own agenda. Lucky for me @cloudflare told them to jog on.

  • AjaySohmshetty
    Ajay Sohmshetty (@AjaySohmshetty) reported

    For context- Cloudflareโ€™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 โ€œstepsโ€ - 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 warningโ€ฆ Feeling blindsided after weโ€™ve already fully built all of our durable workflows on Cloudflare

  • devscipline
    Altin (@devscipline) reported

    @AjaySohmshetty @Cloudflare @andrewk17 And that is precisely why I never put my billing details on any account that does not have some minimal billing limiting settings. I don't like to get bankrupt just because they didn't do their work.