Deploying applications to more Availability Zones than necessary will linearly increase the availability and reliability of the application.

Availability Zones automatically manage the state of applications, enabling stateful applications to fail over without any data loss or manual intervention.

Applications should rely on a single Availability Zone for mission-critical deployments to concentrate the resources and strengthen the zone's resilience.

Applications must be designed to handle the failure of an Availability Zone without human intervention, automatically redistributing traffic to instances in operational zones.

Using diverse Availability Zones for different application components eliminates the need for synchronous data replication.