Tell HN: Cloudflare silently injects its analytics when you switch nameservers
A few hours ago I switched my nameservers to Cloudflare in order to enable R2 bucket serving through my own subdomain, and I found out that it silently had injected a JS analytics snippet in my HTML-only JS-free site textlog.cc — I had to go to the Analytics dashboard, Add the site to the analytics and then disable the snippet. I find this approach entirely invasive, you should opt-in to features like that not have to opt-out. Just a warning out there to folks who might not be aware of this.
You are right that Cloudflare enabled these analytics by default for our free plans in Septemeber of last year.
We built Real User Measurement (RUM) into our free plans because it gives site owners actionable performance data they would not otherwise have. It is on by default for free sites fr the reasons we wrote about in the blog post below. It is easy to disable if you don't want it on. All of our paid plans are opt-in only.
This also gives free plans access to our Observatory product at no cost. Observatory is a performance-monitoring tool inside the Cloudflare dashboard that combines real user data with simulated lab tests to help you measure and improve your website speed.
This makes the client only load self-hosted scripts, or scripts only from the specified origins, among the other directives CSP allows (e.g. restricting styles, images, frames, etc.): https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/CSP
If Cloudflare (CF) has r/w access to the response body, which CF does have by default, then CF can easily modify or remove that <meta> tag. The risk is not abated
The risk of third parties injecting scripts, etc., e.g., analytics, advertising, etc., into response bodies (web pages) is usually cited as a rationale for using HTTPS^1
CF somehow avoids the usual objections. CF is a MiTM but few people object
1. For example, a data collection, surveillance and advertising services company that operates a www search engine and releases a web browser may not want an ISP to inject scripts, etc., e.g., analytics, ads, etc., into web pages as it might compete with the company's business. As a defense against such ISPs and other third parties that are potential competitors for data collection/surveillance/advertising services, it might favor HTTPS sites in its www search engine results, promote HTTPS at conferences discussing its web browser, etc.
FWIW, I operate own DNS (including own custom root.zone) and I MiTM own TLS traffic with a localhost forward proxy. With this setup I get r/w access to response bodies, I add a CSP as an HTTP response header, and a long list of other traffic manipulation. There is no tracking, ads, telemetry, etc. Nothing leaves the computer unless I allow it. Operating DNS plus forward proxy gives me lots of control
Letting Cloudflare (CF) operate DNS and direct traffic through its proxies gives CF control
It's interesting to see how they use it under market pressures
Thanks for the background context, very interesting and realistic. Yes, of course it also has read/write access, though I always make sure there’s no interference of that kind. It has only happened to me once, where a binary I hosted on Pages didn't work when downloaded with wget (though that happened several years ago).
Yup, I explicitly had all anaytics turned off. But had a few sites using Cloudflare for caching. Now I'm checking and seeing this on all of them. This is gross and unacceptable. "Caching" does not mean "modifying my site".
I'm not seeing this on my site (if you want to check: https://stackgho.st), are you using their `strict` http settings? I.e. is your server terminating TLS or is theirs?
edit: perhaps it's only for sites added after that policy came into effect
Cloudflare sets up a reverse proxy as part of their core offering, so by default they can MITM your proxy. The “orange cloud” by a DNS record means it points to their proxy instead of your server.
This reminds me of the old free hosts that would inject their own scripts/ads on pages hosted by them, but their implementation would do it only after detecting the closing HTML tag (either </body> or </html>, can't remember) and the various tricks to get around it, including omitting the tag (browsers don't care), rAnDoM case tags (some were hardcoding lower or upper and comparing case-sensitively), and messing with the content type and relying on the browser to detect correctly since it didn't inject into other types (not sure how the current browsers will handle that).
Are you using CF as a proxy or only for DNS? I ask because I just went to check my domains on the dashboard (some purchased a few years ago, one purchased just a couple days ago), and none of them have Web Analytics enabled.
I have all my domains set to DNS only, so no CF proxy. Wondering if that is why?
I just went and switched them to DNS only, they were on Proxy by default, that also should have been opt-in. Maybe I missed the option while switching the nameservers because I did it in a hurry but still.
I agree it is a bit confusing, but it makes sense if you think about the expected default use case, which is enabling the cloudflare CDN for customers. They expect people to be using their CDN.
I use them for my registrar, because they offer domains at cost and you can’t beat that. But yes, they make it an extra step to use them without using the CDN.
Once you figure it out, it is fairly straightforward, though, and you can’t beat the price.
A simple screen when I did the switch that requires explicitly ‘Enable Proxy’ or ‘skip for now’ would have been enough. Also for ‘Enable Analytics’. None of these existed, or not very obvious otherwise I’d seen them.
Yes, they add the js if "web analytics" is enabled. I believe I had to manually enable it on my old sites though. Maybe it's enabled by default when adding new domains?
When you set up CNAME and certain other records in Cloudflare DNS, it defaults to (and heavily discourages you to disable) "proxied" records, which I believe means that the record points to a Cloudflare-owned host which then acts as a reverse proxy to whatever value you'd set. So from the console it looks like you've set the CNAME to a certain value, but in practice it'll be set to a different thing and transparently forward everything via Cloudflare. This is probably where the analytics get inserted, alongside a bunch of other Cloudflare features.
You can disable this, at which point the record will be set as a normal DNS record.
I can see the advantage of Cloudflare's proxy systems, but I wish they'd be clearer about when they're being used and not pretend that this is some DNS feature or that records have been set to one thing when they've actually been set to something else. If nothing else, it makes debugging DNS issues a lot more confusing, particularly if you're not a DNS expert.
>I can see the advantage of Cloudflare's proxy systems, but I wish they'd be clearer about when they're being used and not pretend that this is some DNS feature or that records have been set to one thing when they've actually been set to something else. If nothing else, it makes debugging DNS issues a lot more confusing, particularly if you're not a DNS expert.
You could say the same about the reverse, ie. people set up their site on cloudflare, thought it was "protected", but really it's dns only and their servers are wide open. It's even worse if they migrated from another provider that was providing ddos protection.
> It seems incredibly unlikely cloudflare does this when just DNS hosting
Not to mention impossible when ‘just’ DNS hosting. Though I suppose they could secretly replace the stated IP with one of their own anyway and then still proxy the content.
Yes, the person who controls the DNS controls the certificate.
What a certificate is supposed to verify is that traffic is going to the right place. If you designate cloudflare as the rightful host of your website then they can get a certificate.
This isn't an edge case though. This is cloudflare's primary product. It is why users use them.
They have a product called Magic Transit that offers DDoS protection and such for plain IP traffic, where Cloudflare does not terminate TLS. Pricing is not public but starts in the five-digit USD per month range according to people I talk to.
This may tell you something about how keen Cloudflare are to handle traffic they themselves cannot decrypt.
I can’t recall if there was a setting to enable reverse proxy, if there was it was On by default since I didn’t expect to have reverse proxy enabled as well. But you can also rp without injecting a script. That’s overdoing it.
I don't know what happened in this situation but beware that CF and similar providers are not true DNS hosting providers. They do DNS, but only so their CDN stuff works, and to lock their customers from using whatever DNS hosting they want. Various things that one might reasonably want to do with your DNS zone are not possible with their product. So use it only because you need to do so in conjunction with their core services.
TLS terminates at Cloudflare, not at your origin. When a record is proxied (the orange cloud), CF holds the certificate the browser validates against and opens a separate connection to your server, so it sees plaintext on both sides and can rewrite the HTML on the way out. Same mechanism that makes the WAF and caching work, so it isn't specific to the analytics feature.
Worth checking which of your records are actually proxied. DNS-only ones (grey cloud) pass straight through and can't be touched.
I noticed the same thing with email-decode.min.js on my site. It turns out it's the "Email Address Obfuscation" feature, which I didn't expect to be on by default.
Noticed this the other day as well. Sketchy as fuck. I didn't have analytics enabled. I had to go and enable to get access to the option to turn this off
A public service reminder to anyone setting up their domains that there are many alternatives to Cloudflare, like EG Bunny.net (I am not affiliated, just a happy customer) and that by using one of these other DNS providers, you
- play your small part in keeping the Internet open
- delay or avoid entirely the enshittification that is 100% foreseeable when buying the "free" services of a quasi-monopoly vendor.
If you're not an enterprise customer, alternate providers will also give you every single feature you could ever need and more, and fewer surprises like the OP's.
No it's not. You don't have to like them, but malware is "any program or file built to damage, disrupt, or secretly gain access to a computer, phone, or network". Cloudflare is many things, some good, some bad, but they don't secretly gain access to your system, you have to choose to install it, so it's not malware.
We use Cloudflare tunnels for connecting distributed workers to central infrastructure. Haven't seen this on tunnel traffic, but good to know it happens on nameserver-managed sites. Another reason to audit what your CDN injects — same applies to checking your security headers regularly.
Oh gosh I didn’t enable anything like that also. I just wanted the nameservers in order to serve the bucket under my subdomain. What else is there I wonder?
The main reason anyone chooses cloudflare is for their CDN (which I suppose is a "reverse proxy"...not a way I'd refer to it but technically yes that's essentially what it's doing). If you don't need the CDN or any of the other fancy features there are plenty of straight DNS providers out there (and often provided by the registar these days).
OK, how can Cloudfare edit your HTML without it passing through your server?
If the browser connects directly to your web server, how can there be Cloudfare's analytics stuff?
Check what IP address you are connecting to when you load textlog.cc. Is that an address that you control? If it's not an address that you control, where is it getting your page, and is that not called proxying?
"can't" is incorrect. As your name server, they can decide to turn the reverse proxy/caching at any point in the future, write an obscure blog post about how this is the best solution for all (a post that gets 2 comments on HN due to other news that week), and you'd only find out about it 11 months later in a thread like this one.
Surprise! The man in the middle man-in-the-middles!
This is only the beginning, when you’ll get used to this they’ll do worse and worse, enshittification, remember?
Cloudflare is doing this already. Once they had enough monopoly power, they started a program to block all bots that don't undergo invasive KYC procedures. Eventually, they might become a KYC broker for regular browser users too. The free internet is over.
We built Real User Measurement (RUM) into our free plans because it gives site owners actionable performance data they would not otherwise have. It is on by default for free sites fr the reasons we wrote about in the blog post below. It is easy to disable if you don't want it on. All of our paid plans are opt-in only.
This also gives free plans access to our Observatory product at no cost. Observatory is a performance-monitoring tool inside the Cloudflare dashboard that combines real user data with simulated lab tests to help you measure and improve your website speed.
Blog post: https://blog.cloudflare.com/the-rum-diaries-enabling-web-ana...
This makes the client only load self-hosted scripts, or scripts only from the specified origins, among the other directives CSP allows (e.g. restricting styles, images, frames, etc.): https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/CSP
The risk of third parties injecting scripts, etc., e.g., analytics, advertising, etc., into response bodies (web pages) is usually cited as a rationale for using HTTPS^1
CF somehow avoids the usual objections. CF is a MiTM but few people object
1. For example, a data collection, surveillance and advertising services company that operates a www search engine and releases a web browser may not want an ISP to inject scripts, etc., e.g., analytics, ads, etc., into web pages as it might compete with the company's business. As a defense against such ISPs and other third parties that are potential competitors for data collection/surveillance/advertising services, it might favor HTTPS sites in its www search engine results, promote HTTPS at conferences discussing its web browser, etc.
Letting Cloudflare (CF) operate DNS and direct traffic through its proxies gives CF control
It's interesting to see how they use it under market pressures
<script type="module" src="https://static.cloudflareinsights.com/beacon.min.js/v4513226..." integrity="sha512-ZE9pZaUXND66v380QUtch/5sE9tPFh2zg45pR2PB0CVkCtOREv2AJKkSidISWkysEuQ0EH8faUU5du78bx87UQ==" data-cf-beacon='{"version":"2024.11.0","token":"c0859b51a7804ab5a9cc8e9e2b2c4cde","r":1}' crossorigin="anonymous"></script>
Good point. There should be a place where hackers can get this kind of news. Some kind of Hacker News place.
edit: perhaps it's only for sites added after that policy came into effect
The RUM Diaries: Enabling Web Analytics by Default
(2 points, 11 months ago): https://news.ycombinator.com/item?id=45291323
(1 point, 11 months ago): https://news.ycombinator.com/item?id=45339321
You must be allowing Cloudfare to terminate your HTTPS connections; i.e. using them for actual proxying.
But DNS/Name servers do not see HTML traffic, as the above poster mentioned, so it's not obvious how this change would matter.
Is it that they are serving their HTML via CloudFlare, and cloudflare is making changing in its serving of their html?
I have all my domains set to DNS only, so no CF proxy. Wondering if that is why?
I use them for my registrar, because they offer domains at cost and you can’t beat that. But yes, they make it an extra step to use them without using the CDN.
Once you figure it out, it is fairly straightforward, though, and you can’t beat the price.
The proxy is like 99% of why people use Cloudflare
Cloudflare injected hostile code into a site they are not even hosting? If it's HTTPS, how do they even do that?
Does it violate the "exceeds authorized access" provision in the Computer Fraud and Abuse Act?
It seems incredibly unlikely cloudflare does this when just DNS hosting, if for no other reason then that this would break so many things.
You can disable this, at which point the record will be set as a normal DNS record.
I can see the advantage of Cloudflare's proxy systems, but I wish they'd be clearer about when they're being used and not pretend that this is some DNS feature or that records have been set to one thing when they've actually been set to something else. If nothing else, it makes debugging DNS issues a lot more confusing, particularly if you're not a DNS expert.
I think a better question is: why would you be using cloudflare, if you didn't want that?
You could say the same about the reverse, ie. people set up their site on cloudflare, thought it was "protected", but really it's dns only and their servers are wide open. It's even worse if they migrated from another provider that was providing ddos protection.
Not to mention impossible when ‘just’ DNS hosting. Though I suppose they could secretly replace the stated IP with one of their own anyway and then still proxy the content.
Or is it that when you sign over DNS to a provider, they can take over your cert? They can "ass-cert" their own? :)
What a certificate is supposed to verify is that traffic is going to the right place. If you designate cloudflare as the rightful host of your website then they can get a certificate.
This isn't an edge case though. This is cloudflare's primary product. It is why users use them.
This may tell you something about how keen Cloudflare are to handle traffic they themselves cannot decrypt.
Realistically its a totally different product, and 5 digit price is probably cheap relative to competitors in that space.
Worth checking which of your records are actually proxied. DNS-only ones (grey cloud) pass straight through and can't be touched.
https://limereader.com/
Make a CNAME record the same way you would for a CDN subdomain.
(I am not yet running this in production, YMMV.)
It this point in time it is somewhere between GoDaddy and RyanAir in dark pattern usage
- play your small part in keeping the Internet open
- delay or avoid entirely the enshittification that is 100% foreseeable when buying the "free" services of a quasi-monopoly vendor.
If you're not an enterprise customer, alternate providers will also give you every single feature you could ever need and more, and fewer surprises like the OP's.
If they can inject script, they can also snoop on all your cleartext traffic without you knowing....
If the browser connects directly to your web server, how can there be Cloudfare's analytics stuff?
Check what IP address you are connecting to when you load textlog.cc. Is that an address that you control? If it's not an address that you control, where is it getting your page, and is that not called proxying?
[1] - https://www.youtube.com/watch?v=a3Xxi0b9trY
And they can fuck off.