Datamover failover is NOT triggered by the loss of all ethernet ports. If you want switch redundancy, you can use the FailSafeNetwork feature and connect each leg of the FSN to a different switch. If you have newer network equipment that can support etherchannel or LACP trunks across multiple switches, that can also be used to provide network redundancy.
A datamover failover occurs as a result of a datamover panic which can be caused by hardware failures, loss of backend storage, and other similar events.
The only datamover failover settings to my knowledge adjust whether failover occurs automatically or waits for administrator, and there is an option to attempt a datamover reboot before failing over as well.
If the physical port hardware inside the datamover fails, it is unlikely to trigger a datamover failover. Redirection of FailSafeNetwork is less disruptive to users compared to a datamover failover.