I am researching using Kubernetes on baremetal for our application. There's an issue I can't find info about. If a worker node dies, how does Kubernetes determine which node(s) to redirect its workload to? What I'm mostly worried about is that if a node dies due to being overloaded with work, Kubernetes will redirect its workload and another node will die due to the workload increase, which leads to a domino effect of nodes dying. How does Kubernetes prevent this?
In order to avoid this kubernetes reserves resources on the nodes for its daemons. See full explanation and configuration guidelines here