The delay of the second database reconnect should be reduced

Current Situation

• Currently we face the situation that the JobScheduler tries immediately to reconnect to a database if the connection gets lost.
• If the database is not accessible then the JobScheduler tries to reconnect every minute.
• If the JobScheduler runs in a cluster then the second database reconnect is too late.
  • If the database is not accessible while the JobScheduler tries to write a heartbeat into the database and the first reconnect fails then the last successful heartbeat lies 2 minutes back.
  • If the JobScheduler tries the second database reconnect and is not successful then the JobScheduler shuts down.
• For example, it is possible that the JobSCheduler in a cluster shuts down because of a database server restart

Desired Behavior

• The second reconnect is performed earlier (after 40 seconds).
• If the database is down for a short time then the JobScheduler in a cluster doesn't shut down.