Differentiate high availability (within a region/cluster: multiple workers, clustering, persistent queues) from disaster recovery (cross-region failover, RPO/RTO) and select the correct approach for each deployment option.