The Vertical Pod Autoscaler (VPA) is an open-source Kubernetes component that automatically sets the CPU and memory requests of containers based on measured usage. Pods with over-generous requests are scaled down and pods with requests that are too tight are scaled up. You configure it with a VerticalPodAutoscaler resource and install it from the kubernetes/autoscaler repository, which also provides a Helm chart. The repository lists the VPA as beta. It is an open alternative to commercial right-sizing services for Kubernetes that adjust requests based on usage data.
The VPA is developed by Kubernetes SIG Autoscaling and licensed under the Apache License 2.0. Kubernetes is hosted by the CNCF, part of the Linux Foundation, which is based in San Francisco (US). You run the component yourself in your cluster; there is no hosted service. Usage data from your workloads therefore stays inside your own environment.