(I also asked the question on SO, and got a NO answer. But before giving up on k18s (since I don’t really want to give up the nice features k18s offers) I’d like to ask the community again just to confirm.)
I have a bunch of servers, geographically dispersed, that are used to let users ping an IP they specify to test network latency from different parts of the internet.
I wonder if it’s appropriate that I use k18s to manage them? Specifically, I want to quickly bring up servers in all regions with the docker image I have, and I want to make sure there is at least one pod in each region that is live. Servers don’t necessarily have to talk to each other, but they should be able to use some shared config.
From the docs it seems Kubernetes assumes nodes should live close to each other the cluster shouldn’t be dispersed.
Should I use k18s for such task or I should consider some other options?
There are many things that just use lot of communication over the network (between nodes and control plane and even kube-proxy). A cluster is not really designed to be geographically distributed right now :-(.
If having multiple clusters is an option, I think you can consider it. But, is it an option?
There won’t be many servers in each region, but there will be many regions. Not sure if multiple clusters can bring much benefit vs just mange each server manually.