Fully agree with this. I have used GKE at scale as an example. It starts to break down significantly once your cluster reaches a couple hundred nodes. Only way

Fully agree with this. I have used GKE at scale as an example. It starts to break down significantly once your cluster reaches a couple hundred nodes. Only way to solve it is to have support resize your controller nodes manually. And as support for anything google is horrible, that task takes multiple hours.