A company runs an event-ticketing platform on AWS and wants to lower costs. The platform uses Amazon EKS on Amazon EC2 for most workloads and an Amazon RDS for MySQL DB instance. New features will run on Amazon EKS with AWS Fargate. The workload has infrequent, date-driven demand spikes. Which approach is the MOST cost-effective for this environment?

The correct option is the most cost-effective because it combines several AWS cost optimization strategies tailored to the described workload. Compute Savings Plans offer significant savings (up to 66%) over On-Demand pricing for a consistent compute usage commitment, making them ideal for the EKS cluster's baseline. On-Demand Capacity Reservations guarantee capacity for predictable, date-driven spikes, ensuring availability without over-provisioning long-term. No Upfront Reserved Instances for the database provide savings for the base load without a large initial payment. Scaling read replicas during peaks offloads read traffic, improving performance and cost-efficiency compared to scaling the primary instance. The other options are less optimal. Using Standard Reserved Instances for EC2 and Spot Instances for peaks is good but less flexible than Compute Savings Plans, which cover Fargate and EC2. Buying a 1-year All Upfront RI for the database sized for peak load is expensive if the peak is infrequent. EC2 Instance Savings Plans are more restrictive than Compute Savings Plans. Manually scaling up the DB instance temporarily for peaks is less efficient and potentially more disruptive than read replicas.