What is the MOST cost-effective way to resize the cluster?
Decrease the number of nodes in the ElastiCache for Redis cluster from 2 to 1.
Deploy a new ElastiCache for Redis cluster that uses large node types. Migrate the data from the original cluster to the new cluster. After the process is complete, shut down the original cluster.
Deploy a new ElastiCache for Redis cluster that uses large node types. Take a backup from the original cluster, and restore the backup in the new cluster. After the process is complete, shut down the original cluster.
Perform an online resizing for the ElastiCache for Redis cluster. Change the node types from extra-large nodes to large nodes.
Explanations:
Reducing the number of nodes from 2 to 1 would compromise high availability, which is a requirement for the application. This is not a cost-effective solution for maintaining availability.
Deploying a new ElastiCache cluster with large nodes and migrating the data is unnecessary and more complex than needed. This involves downtime and is not the most cost-effective way to resize.
While deploying a new cluster and restoring a backup might work, it introduces unnecessary complexity, downtime, and extra costs compared to online resizing.
Performing an online resizing by changing the node type to a smaller, more cost-effective size (large nodes) is the most cost-effective way to resize without impacting high availability. ElastiCache supports online resizing for node type adjustments.
I systematize that the answer is:
Perform an online resizing for the ElastiCache for Redis cluster. Change the node types from extra-large nodes to large nodes.