Mirth hang when backend DB failedover

[elavy-harris]

We had a case where our connected back end Azure Postgres Flexible server did a failover and mirth hung. After investigation, it seemed that some of the DB connections mirth had established still pointed to the old node that was read only, but mirth did not realize the connection was read only.

It should be noted that we are using kubernetes to run connect pod in a single, non HA pod. The image below shows the connections within the connect container where it has connections to both nodes of the backend azure database

Is there a way for mirth to try a write on its connection to make sure its still writeable? Or is there a configuration to suggest to ensure this doesn't happen.

Or any other suggestions?