(K8s 3/4) Sweep expired suspended Kubernetes environments

What does this MR do?

The Kubernetes provider deletes the runner's PVCs that are older than the TTL plus 10 minutes and that no pod mounts.

Why was this MR needed?

An environment that the orchestrator never releases holds its storage forever.

What's the best way to test this MR?

go test -race ./executors/kubernetes/... -run Sweep

What are the relevant issue numbers?

Related to #39469.