DevOps Glossary

Kubernetes PriorityClass is a cluster-level Kubernetes object that assigns scheduling priority to pods, so the scheduler can decide which workloads should run first when CPU, memory, or other resources are limited. A PriorityClass defines a name and a numeric value, and pods reference it through priorityClassName; higher values mean higher priority. If the cluster cannot place a high-priority pod, Kubernetes may preempt lower-priority pods by evicting them to free resources. Teams use PriorityClasses for critical workloads such as DNS, monitoring agents, ingress controllers, payment services, or batch jobs with strict deadlines. They help protect important services during resource pressure, but they must be used carefully because overusing high priorities can cause less critical workloads to restart often or fail to schedule. For example, a production API pod might use a higher PriorityClass than a development test job, so the API stays available when the cluster is under load.