A company runs a SageMaker-hosted recommendation model in a single Availability Zone and requires an RTO (recovery time objective) of 5 minutes for high availability. Which option achieves high availability with the least effort?

Deploying multiple instances of each production endpoint across at least two subnets in a second Availability Zone within the same Region achieves high availability with the least effort. This strategy provides redundancy within the same region, protecting against single Availability Zone failures and meeting the RTO. SageMaker handles the underlying infrastructure for these deployments. Deploying across multiple AWS Regions is more complex and incurs higher latency and cost, making it unnecessary for a 5-minute RTO if a regional outage is not the primary concern. SageMaker automatic scaling helps with performance under varying load but doesn't inherently provide high availability against Availability Zone failures. Regularly backing up and deploying in a second region is a disaster recovery strategy, not a high-availability solution for a 5-minute RTO, as manual deployment introduces significant recovery time.