Renew EC2 Reserved Instances and add a Compute Savings Plan for serverless

Why the Answer Is Correct

The company already knows its EC2 shape, because it bought Standard Reserved Instances three years ago for the same workload it is running today, and it needs that capacity for three more years. Renewing those Reserved Instances with an All Upfront three-year term gives the deepest discount available for a known, steady EC2 baseline. The new serverless workload is the reason a second instrument is needed, because Reserved Instances do not apply to Lambda or Fargate charges at all, while a Compute Savings Plan discounts EC2, Fargate, and Lambda usage. Purchasing the Compute Savings Plan in the management account lets it cover the additional compute costs across the organization, and the All Upfront term for three years maximises the discount. Between them, the EC2 baseline is covered at the Reserved Instance rate and the serverless and incremental compute spend is covered by the plan.

Why the Other Options Are Wrong

B: A one-year Compute Savings Plan with No Upfront payment gives a smaller discount than a three-year term with an upfront payment, and buying it in each member account rather than centrally provides no organisation-wide coverage, so this is the most expensive of the strategies for a three-year commitment. C: An EC2 Instance Savings Plan applies only to EC2 instance usage and cannot cover Fargate or Lambda charges, so the new serverless workload would be entirely uncovered. No Upfront payment over three years also gives a smaller discount than All Upfront, and the plan is described as covering EC2 in each Region from the management account, which is not how Instance Savings Plans aggregate. D: An EC2 Instance Savings Plan per member account neither covers the serverless workload nor aggregates across accounts, and No Upfront-style discounts on a per-account basis are weaker than a centralised three-year purchase.

Community Comment Notes

The community voted 100 to 0 for A, and the top-voted comment stated the reasoning cleanly, that with no constraints other than cost in the problem statement the correct move is simply to select the cheapest plan, and that Reserved Instances for EC2 plus a Compute Savings Plan for serverless is the right pairing. Another noted the serverless mention is effectively a distractor because Compute Savings Plans cover it while Reserved Instances cannot.