The single-region high-availability (HA) architecture is an enterprise-level production architecture and the most utilized deployment model.
The single-region high-availability architecture is robust and resilient within the scope of a single region. It is the recommended setup, which assumes that all components are actively available across all three availability zones within a cloud region. Databases are clustered within a single region and a single availability zone failure does not affect the stack functionality
The service can be unavailable if all three AWS availability zones are unavailable within the AWS region.
Multi-region active-active is an enterprise-level architecture that adds exact replicas across various regions to serve a regional traffic with the region-based traffic affinity. It is recommended for global production deployments where the services needs to span multiple AWS regions within a geographical location.
The multi-region active-active architecture involves complex tasks, such as monitoring the real-time replication status, correcting replication issues (and similar) for datastores across regions. This setup warrants a heavy monitoring and a constant operational upkeep of the systems involved.
In this mode, there is no need for a separate failover environment. The infrastructure cost of operations and monitoring is high on this kind of configuration.
Multi-region active-active is a technically-advanced multifold architecture, which might translate into time required for its maintenance and maintenance-related costs.