Quick Answer

The answer is configuring a VM Scale Set with a minimum instance count of 2 and spreading across fault domains, as this ensures high availability by distributing workloads across isolated hardware boundaries within an Azure datacenter. This approach protects against both hardware failures and planned maintenance events because fault domains separate VMs onto different physical racks with independent power and networking, while update domains stagger reboots during patching. On the AZ-104 exam, this concept tests your understanding that high availability requires redundancy at the infrastructure layer, not just a single VM; a common trap is assuming a single VM in an Availability Set is sufficient, but you must have at least two VMs to qualify. Another valid method is deploying VMs across Availability Zones, which protects against entire datacenter failures by placing instances in physically separate locations within a region. The Azure Load Balancer then distributes traffic across these healthy instances, but it is a traffic-routing tool, not a high-availability method itself. Remember the memory tip: “Two VMs, two domains, or two zones” — if you lack a minimum of two, you lack high availability.