DevOps Glossary

Kubernetes VerticalPodAutoscaler (VPA), or Vertical Pod Autoscaler, is a Kubernetes autoscaling component that recommends or updates pod CPU and memory requests based on actual workload usage. It helps your pods ask for more accurate resources, which improves scheduling and can reduce wasted capacity or resource pressure in a cluster. VPA is useful for workloads with changing or hard-to-predict resource needs, such as batch jobs, data processing services, or applications that were deployed with rough request values. It works by observing usage over time, generating recommendations, and, depending on its mode, either reporting those recommendations or applying them by recreating pods with updated requests. VPA differs from Horizontal Pod Autoscaler because it changes the resources assigned to each pod rather than changing the number of pod replicas, and teams often use both carefully depending on the workload.