Metrics Server

Metrics Server is a Kubernetes component that collects resource usage data in the cluster and makes it available to other Kubernetes components via the Metrics API. Horizontal Pod Autoscaling and Vertical Pod Autoscaling work based on data from Metrics Server. You can view metrics using kubectl top. Node group autoscaling on cloud servers can be configured using Cluster Autoscaler. To monitor, see logs in the Managed Kubernetes cluster or use dedicated applications, for example, Prometheus.

Metrics Server is installed by default in Managed Kubernetes clusters of version 1.27 and higher. You can upgrade the cluster version on a cloud server.

Read more about how Metrics Server works in the article Kubernetes Metrics Server in the Kubernetes documentation.