Availability Zones vs Availability Sets

High availability keeps an application running even when individual components fail. Availability zones protect against a datacenter failure by placing resources in physically separate zones within the same region. For example, you deploy virtual machines across two or three zones, and a Standard Load Balancer distributes traffic to healthy instances in any zone. If one zone fails, the other zones continue to serve traffic. The design tradeoff is that cross-zone traffic adds a small amount of latency, and zonal resources cost more than single‑zone ones.