Availability Zone

An availability zone (AZ) is an isolated data centre within a cloud region. AZs in the same region share low-latency private network links but use physically separate buildings, power, cooling, and uplinks, so a fault in one AZ does not propagate to the others. AZs are the cloud's unit of fault isolation below the region level.

Why multi-AZ matters

• Hardware failures. A single AZ can lose power, network, or cooling. Workloads spanning multiple AZs survive.
• Software upgrades. Provider rollouts often happen one AZ at a time; multi-AZ deployments stay available during these.
• SLA tiers. Many managed services only offer their stronger SLA when configured multi-AZ.

Common multi-AZ patterns

• Load-balanced stateless services. Run replicas across AZs behind a load balancer; if one AZ fails, traffic routes to the others.
• Multi-AZ databases. RDS Multi-AZ, Aurora, MongoDB Atlas replica sets, ElastiCache replication groups.
• AZ-aware Kubernetes. Pod topology spread constraints place replicas in different AZs.

Cost considerations

Inter-AZ traffic in AWS is billed per GB in each direction (about USD 0.01 per GB in 2026). Chatty multi-AZ deployments can produce surprising data-transfer bills; chatty internal traffic is often the largest hidden cost in AWS architectures.