[DBM][postgresql] Agent fails to reconnect to RDS Aurora after failover

[seanamos]

Agent Version: 7.52.1 (this issue has existed for many versions)

Steps to reproduce the issue:

1. Setup a RDS Postgres Aurora cluster with a least one read replica.
2. Run the Datadog Agent in a docker container and ensure it is setup to connect to the cluster's writer DNS address.
3. Manually failover the cluster.

Describe the results you received:
The agent is unable to reconnect to the writer DNS address. It continually reports in its log attempting to reconnect to the correct DNS address, but is unable to do so. The resolved IP address reported in the error log is the stale writer IP.

Manually restarting the agent's container resolves the issue.

Describe the results you expected:
The agent is able to reconnect to the promoted writer without manual intervention.