**kops** helps you create, destroy, upgrade and maintain production-grade, highly available, Kubernetes clusters from the command line. Currently if the worker node dies, pods have nowhere to reschedule to, which has been causing site cyclical outages.