A critical stateful web application runs on two Linux EC2 instances behind an ALB with an RDS for MySQL database. DNS is in Route 53. The resilience objectives are: application tier RPO 2 minutes and RTO 30 minutes; database tier RPO 5 minutes and RTO 30 minutes. The company does not want major architecture changes and wants optimal latency after failover. Which solution meets these objectives?

The correct solution uses AWS Elastic Disaster Recovery (DRS) for the EC2 instances, which continuously replicates data to a recovery Region, meeting the RPO of 2 minutes and RTO of 30 minutes for the application tier. A cross-Region RDS read replica ensures the database RPO of 5 minutes and RTO of 30 minutes. Deploying an ALB in the second Region and using AWS Global Accelerator with both ALBs as endpoints provides optimal latency after failover by directing traffic to the nearest healthy endpoint and updating DNS. The incorrect options fail to meet the RPO/RTO objectives or provide optimal latency. Amazon Data Lifecycle Manager for EBS snapshots and AWS Backup plans typically have higher RPOs and RTOs compared to continuous replication for EC2 instances, making them unsuitable for the application tier's 2-minute RPO. Relying solely on RDS automated backups with replication might not meet the 5-minute database RPO for immediate failover. CloudFront is a CDN and does not provide active-active traffic management for optimal latency across regions like Global Accelerator.