This one had my scratching my head a bit as I wasn't quite sure if Kubernetes was the right place to do this.
Depending on your use case, it might make sense to terminate traffic before it reaches your cluster but that may have the effect of filtering traffic to other applications if not done properly.
In this instance, the Kubernetes cluster in question makes use of the NGINX Ingress Controller and as such, honours a whole bunch of flags.
Before we get into the details, let's set up a small example.
We'll pretend our desktop has an IP address is
192.0.2.3 exactly. We want to allow only a network range of 1 single address so that say; our mobile device with the address
192.0.2.2 can't connect but our desktop can.
In CIDR notation, this would be represented as
192.0.2.3/32, with the
32 effectively meaning "Just this one address" instead of any other devices on the
192.0.2 range, or broader.
With our address block defined, let's look at an ingress:
- host: example.com
- path: /
Ok, we've allowed our desktop to connect but let's try connecting to this ingress from a device we know isn't allowed, such as our laptop on
> curl https://example.com/
Alright, and now from our desktop at
192.0.2.3/24, which we allowed explicitly:
> curl example.com
Success! We've managed to use nothing but an ingress to block specific traffic but you might wonder, why would I ever use this?
Often, you can't make use of OAuth but you want to protect against random internet traffic so you can explicitly allow known IP ranges.