codeready-toolchain / member-operator Goto Github PK

The idler currently works on K8s built in resources such as Pods, Deployments and ReplicaSets to perform kill/scale-to-zero operations on long running user workloads.

Yet a user workload's custom resource may also control the replicas via scale sub resource (https://kubernetes.io/docs/tasks/extend-kubernetes/custom-resources/custom-resource-definitions/#scale-subresource).

In this case scaling deployments and replica sets to zero won't have any effect on the replicas as the custom resource controller will immediately bring back the replicas. The idler will then kill pods and the scale sub resource brings them back immediately.

A sample of such a workload using the scale sub resource is the Camel K integration custom resource.
https://github.com/apache/camel-k/blob/main/config/crd/bases/camel.apache.org_integrations.yaml#L5911

In case the deployment is owned by a custom resource the idler should also support the scale sub resource to make sure that the user workload is stopped properly.