[FEATURE REQ]: Integrate per-partition automatic failover for single write accounts.

[jeet1995]

Background

In single-write accounts, there is no way for a write operation to failover to a different region in case the server partition sees quorum loss as there is only a single write region. With per-partition automatic failover, the goal is to enable a failover of requests routed to the affected server partition to a different region. At a high-level, the backend elects a read region as the new write region for the affected server partition. The Cosmos DB SDK will use pertinent signals from the backend to mark the affected region for that server partition as unavailable and a new override region to use for that very server partition. 

Scope

Aligned with #3981.