Why are the control-plane nodepools separated into different subnets?

[LeXonJe]

Hi there,

I'm currently digging through this repository because I want to learn how this cluster setup works.

Do I understand correctly that if I, for example, want a 3 control-plane node HA-setup where each node is in a different location (e.g. fsn1, nbg1 and hel1), every control-plane node would be in its own nodepool, thus having its own subnet inside of the private network (e.g. 10.255.0.0/16, 10.254.0.0/16 and 10.253.0.0/16).

Wouldn't this use of subnets be a bit of a waste? Or are there benefits for separating the control plane pools into different subnets that I didn't factor in? In the end, all control-plane nodes still need to communicate, so network isolation is also not a real feature here, right?

I'm still a bit new to the whole "Kubernetes with Hetzner Cloud" topic, so I may just be lacking knowledge. I'm grateful for any hint in the right direction.

Thanks for reading! ^^