plugin for drain decisions

**Describe the feature**
when deciding if a node should be terminates (for example for scheduled maintainance) we want to say "don't do it now if the cluster is in trouble" by for example looking at the current node health (tainted / notread / custom node statuses) across the cluster.

**Is the feature request related to a problem?**
No

**Describe alternatives you've considered**
Built-in logic would not work since these decisions could be very custom.

No files, tests, or entry points are named. Start by locating the code that makes node drain or termination decisions and determine how a plugin could inspect cluster-wide node health; done means custom logic can veto a drain when the cluster is unhealthy without hard-coding one health policy.

Written by the indexing model from the issue text.