Skip single replica deployments

<!-- Please answer these questions before submitting your feature request. Thanks! --> **Is your feature request related to a problem? Please describe.** I have the Descheduler running in dry-run mode on a Cluster with a lot of different deployments to test the functionality. I am seeing a lot of evictions for single instance pods, while they do have a PDB with 0 allowed disruption. Because of these messages, I am reluctant to take the descheduler out of dry run mode. **Describe the solution you'd like** I would like an extra feature where I can set the amount of replica's before it is considered for eviction. **Describe alternatives you've considered** I have looked into following alternatives, but they won't work looking at our infrastructure/way of working - Setting a label on single instance pods, and exclude this label from eviction - Deleting the PDB for single pods, and using the ignore pods without PDB - Filtering by namespace - Filtering by priorityclass **What version of descheduler are you using?** descheduler version: v0.33.0 Extra: I have looked at issues https://github.com/kubernetes-sigs/descheduler/issues/271 https://github.com/kubernetes-sigs/descheduler/issues/786