Improve Bulk Scroller resiliency during DB failover

It seems that if there is a Mongo DB failover where primary host is restarted, the scroller will terminate.  This can be a problem if there is a mass operation in progress.  There are some scenarios where this can happen

1. Customer initiates a mass bulk operation without communicating to Cloud team to allow us to properly scale up DB resources
2. The DB is just overwhelmed and fails over
3. Mongo Atlas infrastructure issues or regular maintenance (os patches,minor upgrades,etc)

Is there a way to improve scroller resiliency so that it does not terminate and the operation needs to be started over again