I have few on-prem clusters and my use case for this feature is to not allow workload to run, until I make sure some specific network routes works.

Without these checks even when cluster communication works, persistent storage might not work but scheduler will put pods with PVs on this node anyway.