[Feature] Introduce cached Kubernetes client

Feature (What you would like to be added): We should use a client cache when querying objects from a Kubernetes cluster. This can be done by using the DelegatingClient of the controller-runtime.

Motivation (Why is this needed?): To save network costs and reduce queries to the Kube-Apiserver.

Approach/Hint to the implement solution (optional): See:

• https://github.com/kubernetes-sigs/controller-runtime/blob/b704f447ea7c8f7059c6665143a4aa1f6da28328/pkg/client/split.go#L44
• https://github.com/gardener/gardener-resource-manager/pull/40/files