Skip to content


Knative Serving provides automatic scaling, or autoscaling, for applications to match incoming demand. This is provided by default, by using the Knative Pod Autoscaler (KPA).

For example, if an application is receiving no traffic and scale to zero is enabled, Knative Serving scales the application down to zero replicas. If scaling to zero is disabled, the application is scaled down to the minimum number of replicas specified for applications on the cluster. Replicas are scaled up to meet demand if traffic to the application increases.

You can enable and disable scale to zero functionality for your cluster if you have cluster administrator permissions. See Configuring scale to zero.

To use autoscaling for your application if it is enabled on your cluster, you must configure concurrency and scale bounds.

Additional resources

  • Try out the Go Autoscale Sample App.
  • Configure your Knative deployment to use the Kubernetes Horizontal Pod Autoscaler (HPA) instead of the default KPA. For how to install HPA, see Install optional Serving extensions.
  • Configure the types of metrics that the Autoscaler consumes.
  • Configure your Knative Service to use container-freezer, which freezes the running process when the pod's traffic drops to zero. The most valuable benefit is reducing the cold-start time within this configuration.
Back to top

We use analytics and cookies to understand site traffic. Information about your use of our site is shared with Google for that purpose. Learn more.

× OK