Redis cluster never removes hosts that are PENDING_DYNAMIC_REMOVAL

*Description*:
When a host is removed from a Redis cluster, and health checks are enabled, then the host will remain in PENDING_DYNAMIC_REMOVAL even after it fails health checks.

A test for this bug can be found (commented out) in this commit: https://github.com/envoyproxy/envoy/pull/14483/commits/cc8729f9bcd219bc8f016f21d90f086e63e02b39

It seems likely that this is also true of the strict DNS cluster type too.

*Repro steps*:
Create a Redis cluster with health checks enabled.
Remove a host from the cluster in a way that causes the health check to still pass.
Check that the host has moved to pending dynamic removal.
Fail the health check.

Expected result:
The host is removed from the cluster.

Actual result:
The host remains in pending dynamic removal state.