Question

A retail company runs a stateless web API on individual Azure VMs behind a Standard Load Balancer in a single region. During peak sales events, traffic spikes unpredictably, and the current fixed set of VMs cannot scale automatically. The architect must redesign the compute tier to automatically scale based on CPU load while also surviving a datacenter-level failure within the region. The solution must minimize management overhead. What should the architect recommend?

Accepted Answer

Deploy a Virtual Machine Scale Set with Flexible orchestration spread across multiple Availability Zones and configure autoscale rules based on CPU metrics.

Answer

Place the existing VMs in an Availability Set with two fault domains and manually add VMs before each sales event.

Answer

Deploy a Virtual Machine Scale Set with Uniform orchestration in a single Availability Zone and enable autoscale based on CPU metrics.

Answer

Migrate the VMs to a single large VM SKU with Azure Site Recovery replication to a paired region for failover.