I am fairly a greenhorn wrt to running k8s, but I have reasonable experience using k8s. I wonder how microk8s scales? I’m running two dozen pods on DigitalOcean’s k8s offering. The pods are somewhat busy, but the hardware it’s running on is not some monster machines (three dozen gigabytes of RAM, 20 VCPUs, 7 nodes).

I wish to set up a microk8s cluster on metal and either migrate (part of) the work load to it or expand our capacity with it. How does microk8s scale? Can I expect it to scale well over a setup like the above and beyond? Does anybody have experience running a cluster of comparable capacity under microk8s? In particular wouldn’t dqlite become a bottleneck (as opposed to etcd)?