I’ve migrated from cloudflare pages to cloudflare tunnels as I wanted to do a little bit more.
I can’t segregate my network as my ISPs router is rather limited, which means no vLANs. Connecting another router would introduce a double nat as they don’t allow bridging. So I’m running my website basically “raw” in a hyperV virtual machine. the website is semi-static and made out of flatfiles, therefore it’s is quite impossible to login into it. as stated before i’m using cloudflare tunnels to expose a nginx server to the interner. what are the chances someone or something (bot) inflataring my network? 100% safety is not possible but how safe am i?
You must log in or register to comment.
Meh. Safeish. Until one of your servers has a zero day.
Everything has some sort of vulnerability, the qestion is will someone be assed to abuse it.(rheotical question)
Not so much will someone be assed about it, it’s whether a script will pick you up your server. There’s a ton of aggregation search engines that scan most IPv4 addresses and list them on what ports are open etc. such as Shodan.io
Like I said, safeish.
As far as i understood it a cloudflare tunneled service should not be visible when port scanning. Or am i completely wrong here? I started using tunnels just so i can avoid opening ports to the internet. I also restricted the access to my services to specific countries.
The only thing a CF tunnel does is protect your home IP. Doesn’t protect the app or server you’re exposing.
Well it does slightly more than just obfuscating your home IP, in that it will also do automatic bot, DDOS prevention, etc…
Nothing will stop a general scan from happening. Especially if it’s a slow scan.
Scans won’t trigger dos/ddos alerts.
tunnels are reverse-portforwarding. ports aren’t open on my network but on theirs.
anyways i moved back on VPS because im not 100% sure what is my ISPs stance lmao. and since i cant have much control with my internal network for now, id rather stay away but i def wanna host at home eventually
Well yeah, that would get your host IP…if they’re doing a general scan of whole ISP IP ranges (Which nothing could really stop, except for a good firewall). But there is much more low-hanging fruit for hackers than to scan tens of thousands of unoccupied subnets.
You’ve already taken a great step by setting up Cloudflare tunnels, as that will obfuscate your WAN IP, but a common mistake I see a lot is having another random device on a network that is perhaps using a DDNS that doesn’t obfuscate A records or something like that.
Basically, just make sure everything that is public/internet-facing is going through CF tunnels and you’re as protected as you can (reasonably) can be - from that angle at least.
Keep in mind though, this just (largely) prevents one vector of attack - through your WAN IP - depending on your set-up, you could (and likely do) have other ways of penetration to get into your network.
I am a big proponent of getting something like a Firewalla to mitigate many other vectors. They’re bit pricey (though for their capabilities relative to other “off the shelf” devices, not really, I suppose) but largely hands-off.
You can’t get rid of the ISP router? Can you contact your ISP to put the router into bridge mode?
Do you have any auth in cloudflare? If so, that mitigates a lot of zero-days. First they have to get past cloudflare, then a zero-day in your nginx.
Connecting another router would introduce a double nat as they don’t allow bridging.
Maybe you should try connecting another router nontheless. Double NAT isn’t nice, but not having VLANs and generally being bound to your ISP router isn’t either. I’m behind a double NAT for some time now and my network still works fine.
If you keep up to date with all the vulnerabilities for the software you use you should be generally pretty safe.
Tbh… It’s just nginx and gunicorn and yeah i keep it updated :)
Run your hosting inside a docker container. If someone were to hack you they, in theory, would only have access to whatever is in that container and not your entire server. Someone with more expertise please feel free to correct me if I’m wrong. I am not an expert in this, just offering what I (think I) know.
Exactly. Even if they make it into your network (which would be somewhat unlikely already, since CF obfuscates your home IP), as long as you don’t run your docker containers in “privileged” mode (or give the containers RW access to important directories on your server), you’d be fine except for the most end-of-world scenarios.
