Question 510
Two tiers of a legacy app use a classic availability set in Azure region eastus; the application tier currently has only one VM in that set and its load balancer has no second backend target. The operator must protect against host failure within the regional availability-set design rather than implement autoscaling. What is the most direct correction?
Explanation
Correct Answer
An availability set spreads multiple VM instances across fault and update domains to reduce correlated maintenance and hardware exposure. With only one VM, no alternate application instance exists when the first fails. Deploying another replica into that same availability set and configuring a suitable health-probed load-balancer backend provides actual instance redundancy. Confirm that application data/session handling supports a second instance and that both VMs can serve requests before relying on the improved availability.
Incorrect Answers
A VM scale set can provide scalable redundant compute, but replacing an established legacy availability-set deployment is not the direct requirement when capacity elasticity is explicitly out of scope. The existing set can accommodate more than one application VM. Adding a redundant replica within that supported topology is the minimal targeted correction.
A load-balancer frontend with only one backend instance does not create a failover path if that VM fails. The frontend can direct traffic only to configured and healthy backend targets, so an additional VM and correct pool membership are required. Merely changing a virtual IP cannot produce application availability.
Management locks control resource changes through Azure Resource Manager; they do not clone VMs, distribute one VM over multiple physical fault domains or protect it from hardware loss. The single-instance application would still have a single compute point of failure. Add a real second VM rather than a control-plane lock.