Add mode for scale down instead of delete upon meeting certain conditions

Some users would prefer to have a mode to scale down the deployment of its associated pod(s) instead of straight-up deletion.

Start by locating the logic that removes Kubernetes resources after the configurable pod restart threshold, then trace how associated deployments and pods are identified. Clarify which conditions should trigger scaling down, how the mode is configured, and what tests cover deletion today. Done means the selected mode scales down the deployment instead of deleting it when those conditions are met.

Written by the indexing model from the issue text.