EC2 it appears does not do this (capture a kernel panic reason and put it somewhere that an orchestration system can find it later). Kernel panics are a thing.

EC2 it appears does not do this (capture a kernel panic reason and put it somewhere that an orchestration system can find it later). Kernel panics are a thing. Whether a kernel is hosting k8s processes seems not terribly relevant to the probability it will panic? k8s seems neither likely to induce panics more often, nor to reduce their probability, imho. This is probably a case of "at scale, everything happens", including kernel panics in VMs that most people assume never panic.