The correct answer is C because the workload is a batch processing job that runs during a limited time window each night and can tolerate interruptions. The key statement is that if a job fails on one instance, another instance will reprocess the job . That means the workload is fault-tolerant and well suited for Amazon EC2 Spot Instances , which provide unused EC2 capacity at a much lower cost than On-Demand Instances.
Batch jobs, big data processing, background analytics, and other interruptible workloads are classic use cases for Spot Instances. Because the instances run only from 12:00 AM to 06:00 AM and do not need to be available continuously, committing to a 1-year Savings Plan or Reserved Instances would not be the most cost-effective approach. Those options are better for steady and predictable long-running usage, not for short-duration nightly workloads.
Using a new launch template configured for Spot Instances allows the Auto Scaling group to launch lower-cost capacity for the processing window. Scaling based on CPU usage can help the group add or remove instances according to processing demand, improving both elasticity and cost efficiency.
Option A is less cost-effective because Savings Plans are more suitable for consistent usage over time. Option B is also less cost-effective because Reserved Instances lock in a commitment for capacity that is used only a few hours each day. Option D does not address the cost model and could even increase cost by using larger instances.
AWS cost optimization guidance recommends Spot Instances for fault-tolerant, flexible, interruption-tolerant workloads , making option C the best answer.
