K8s Cluster Protection — Priority Classes

A Kubernetes cluster may stop working or degrade because a critical component is evicted when the cluster is short on resources, and thus important apps/add-ons become pending scheduling. Applications such as Monitoring, Logging, or Falco are critical from the security perspective and must keep running even in the most challenging situations.

The scope of today’s blog post is to describe how to prevent mission-critical and system workloads from becoming permanently unavailable.