Question

A startup runs a session-caching layer on a self-managed ElastiCache for Redis cluster with two large node types provisioned 24/7. Traffic is very low overnight and on weekends but spikes unpredictably during weekday marketing campaigns. The finance team wants to stop paying for idle capacity while still handling sudden spikes without manual intervention. Which change best optimizes cost for this pattern?

Accepted Answer

Migrate to ElastiCache Serverless, which automatically scales capacity up and down based on demand and charges for the data stored and compute used.

Answer

Purchase Reserved Nodes for the existing cluster to lower the hourly rate of the always-on nodes.

Answer

Add a scheduled scaling policy to remove nodes overnight and re-add them each morning.

Answer

Replace ElastiCache with a self-managed Redis cluster on Spot Instances to reduce compute costs.