Skip to main content

KubeDownscaler

The KubeDownscaler is the main component of GoKubeDownscaler. It is responsible for scaling workloads based on the defined schedules and configurations. It periodically checks the current time against the defined schedules and scales workloads up or down accordingly.

How It Works​

The Downscaler operates in a loop, performing the following steps:

  1. Scan Namespaces and Workloads: The Downscaler scans the specified namespaces and workload types to collect resources that should be evaluated.
  2. Evaluate Each Workload: Each workload is evaluated against the defined schedules and configurations to determine if it needs to be scaled up or down.
  3. Scale Workloads: If a workload needs to be scaled, the Downscaler performs the scaling operation, adjusting the number of replicas as specified in the target configurations.
  4. Repeat: The Downscaler repeats the process, continuously monitoring and scaling workloads as needed.

The Downscaler can be installed and operated cluster-wide or restricted to specific namespaces

The details of how the Downscaler evaluates workloads and determines scaling actions are covered in the Scopes And Scaling section.