Consider supporting scale-down for a cluster with Not Ready members

There is one specific circumstance where we might want to support scaling down a cluster, even though it has members that are Not Ready:

- It is a 5+ Member cluster
- There are 3 Ready nodes
- Two nodes are Not Ready because the user has resource problems in the Kubernetes cluster
- The user would like to scale down to a 3-node cluster

This is a fairly specific failure case, so we are not building it into release 0.3. This issue is being created to track the idea/possible requirement for the future. Do not start work on it at this time.

Ref: https://github.com/etcd-io/etcd-operator/pull/456#issuecomment-5332566251

No files, tests, or entry points are named. The issue records a future requirement for scaling a 5+ member cluster down to three Ready nodes when two members are Not Ready; work is explicitly deferred, so there is no current implementation target or completion test.

Written by the indexing model from the issue text.