CPU & memory utilization

View resource utilization for your cluster's applications and nodes

Overview

Infra App provides real-time information about cluster and workload utilization, so you can identify how resources such as CPU and Memory are being used and reserved across your Kubernetes cluster.

Metrics Server Required

CPU & memory utilization will work out of the box with most clusters. For real-time usage metrics, you'll need to make sure the official Kubernetes metrics server is installed. Many cloud Kubernetes providers such as GKE include it by default.

Workload utilization

In the main Workloads view, Infra App shows the resource utilization of each workload at a glance. Infra App automatically adds up all the resources used by all pods of your workload, so you can view the total footprint of your workload across the cluster:

To view top workloads by utilization, click on CPU Usage or Memory Usage in the table to sort the workloads:

Individual workload utilization

By clicking on an individual workload, you can see more details about its CPU & Memory utilization:

Hint: CPU & Memory Percent is calculated by comparing the workload's usage against (in order):

  • Requests (if set)

  • Limits (if set)

  • Total cluster capacity

To further view a workload's utilization breakdown by pod, the pod table includes the utilization per-pod:

Cluster Utilization

In the cluster view, Infra App shows what percent your Kubernetes cluster's resources are being used:

In this case, the cluster is almost out of CPU requests, meaning that future workloads may not be successfully deployed by Kubernetes.

Node Utilization

In the cluster view, Infra App includes a list of nodes with CPU & Memory utilization included. You can switch between usage, requests and limits to compare different utilization metrics for your nodes.

Clicking on a node in the table, you'll see the node's utilization at a glance:

Lastly, a list of pods on the node view can show you how resources are allocated on this node: