12.2 Dataguard observer going down during the role transition at primary DB failure event

Dataguard observer going down during the role transition at primary DB failure event

What is the expected behavior of the observer during the fast start failover of the primary database in the event of Primary DB failure and role transition of databases.

We are seeing the observer status is intermittent and sometimes it is up and running and most of the times we need to start them manually to reinstate the failed primary DB which is now current standby database, please advise on how we can keep the observer up and running all the time during the failure events.

We can enable the checkbox to restart of observer when observer is down, however EM is having problems recognizing the observer servers. Is there a manual set configure parameter for the observer to restart automatically?