Your company runs a critical web application on Azure Virtual Machines in a single region. Management requires that the application remain available even if an entire Azure datacenter within that region experiences a physical failure, while keeping all resources in the same region for data residency. Which approach should you use to deploy the VMs?