An ElastiCache for Redis cluster consists of two xlarge nodes in different Availability Zones and the monitoring shows approximately 75% of memory is freeable. The application requires high availability to be preserved. What is the most cost-effective method to resize the cluster?

The most cost-effective method is to perform an online resize of the ElastiCache cluster to change the node type from xlarge to large. Amazon ElastiCache for Redis supports online resizing, allowing you to scale down node types without downtime, preserving high availability. Since 75% of memory is freeable, a smaller node type (large instead of xlarge) is sufficient, reducing costs. Reducing the node count from 2 to 1 would eliminate high availability. Launching a new cluster and migrating data or restoring a snapshot involves more operational overhead and potential downtime or dual-running costs compared to an online resize.