Identify underprovisioned and overprovisioned GKE clusters

This page explains how to identify underprovisioned and overprovisioned Google Kubernetes Engine (GKE) clusters. GKE provides insights and recommendations for cost optimization scenarios such as overprovisioned clusters and idle clusters, and reliability improvement scenarios such as underprovisioned clusters. GKE provides corresponding recommendations to scale up, scale down, or delete the clusters. For idle clusters, see Identify idle GKE clusters.

After you verify that the identified clusters would benefit from the recommendation to scale up or down, you can make the recommended change to save costs or increase the reliability of your cluster. If possible, the recommendation includes projected monthly savings or cost. For more information, see Understand cost or savings estimates.

GKE doesn't provide these insights for Autopilot clusters, which incur minimal operational costs because you only pay for the resources that your workloads request. For more information, see Autopilot Pricing.

GKE monitors your clusters and delivers guidance to optimize your usage through Active Assist, a service that provides recommenders that generate insights and recommendations for using resources on Google Cloud. For more information about how to manage insights and recommendations, see Optimize your usage of GKE with insights and recommendations.

Get insights and recommendations for underprovisioned and overprovisioned clusters

GKE surfaces these insights and recommendations in the following locations in the Google Cloud console:

  • Kubernetes Clusters page, in the following locations:
    • In the Kubernetes clusters list, in the Notifications column for the applicable clusters
    • Notification banners on the Clusters page for a specific cluster
  • FinOps hub

The recommendations have the following titles in the Kubernetes Clusters page:

  • Overprovisioned clusters: "Decrease cluster resources to reduce costs"
  • Underprovisioned clusters: "Increase cluster resources to improve reliability"

You can also receive these insights and recommendations through the Google Cloud CLI or the Recommender API, using the CLUSTER_UNDERPROVISIONED and CLUSTER_OVERPROVISIONED subtypes.

Follow the instructions to view insights and recommendations.

After you identify underprovisioned or overprovisioned clusters, see the considerations when rightsizing clusters.

How GKE identifies underprovisioned and overprovisioned clusters

The following table describes the signals that GKE uses for identifying underprovisioned and overprovisioned clusters that can be scaled up or down, and the threshold for each signal. Additionally, this table shows the action we recommend that you take in this scenario.

Subtype Signal Observation period Details Recommendation
CLUSTER_UNDERPROVISIONED CPU or memory usage is high Last 30 days A GKE cluster is underprovisioned when both CPU and memory utilization average greater than 80% during every hour, over the last 30 days. Scale up your cluster to increase reliability
CLUSTER_OVERPROVISIONED CPU and memory usage is low Last 30 days A GKE cluster is overprovisioned when CPU and memory utilization average between 7% and 20% during every hour, over the last 30 days. Scale down your cluster to save costs

GKE doesn't send recommendations for clusters that were created less than 30 days ago.

Understand cost or savings estimates

If possible, GKE's recommendation includes an estimate that projects the monthly cost or savings if you rightsized the cluster. This estimate is derived from the cluster costs over the past 30 days.

Any estimated costs or savings are projections based on previous spending, and are not a guarantee of future cost or savings.

To see these estimates, ensure that you have the required billing.accounts.getSpendingInformation permission to get spending information. For details, see Cloud Billing access.

To get more information about the cost of all of your GKE clusters, including a more granular breakdown based on namespaces and workloads, see Get key spending insights for your GKE resource allocation and cluster costs.

For more information about the costs of running a GKE cluster, see GKE pricing.

Considerations when rightsizing clusters

Before you follow a recommendation to scale up or down a cluster, consider the following:

  • Review the resource utilization of applications running on your cluster to see how they're performing, and if they're using more or less CPU and memory than expected. For instructions, see Analyze resource requests.
  • Batch processing workloads might intentionally maintain high utilization of cluster resources for cost efficiency. If the allocated cluster resources are sufficient for the batch jobs running on the cluster, you don't need to scale up the highly utilized cluster, which was identified as underprovisioned.

Implement the recommendation to rightsize a cluster

Review the following to understand how you can adjust the size of a cluster to better match your resource utilization.

Rightsize an underprovisioned cluster

To implement the recommendation to minimize the risk of reliability by rightsizing an underprovisioned cluster, increase resources on the cluster. You can do so by taking some of the following actions:

When you implement this recommendation, you ensure that your cluster remains reliable because it has the appropriate amount of resources for its applications.

Rightsize an overprovisioned cluster

To implement the recommendation to save costs by rightsizing an overprovisioned cluster, decrease resources on the cluster. Adjust cluster CPU and memory allocations to match your workload needs. You can do so by taking some of the following actions:

When you implement this recommendation, you ensure that you're not using more resources than necessary to run your cluster's applications.

What's next