A resilient VM design separates several concerns. Placement across failure domains helps availability. Persistent disks and retained backups support recovery. Encryption protects data under its stated key and access conditions. None of those controls replaces all the others.
Read the failure described in each original scenario and choose the evidence or protection that addresses it. A resource-group move is a scope change, not a guarantee of a new physical location. A larger VM doesn't prove that its data can be recovered. These are checked design decisions. The simulator doesn't perform a real disk restore, migrate a VM or test a production availability claim.
Match the recovery control to the actual failure.