Introduction

The AWS Enterprise Support plan offers a comprehensive set of features to help you get the greatest value from AWS. These features include proactive planning, advisory services, automation tools, and 24/7 support. Cost Optimization guidance is a fundamental strategy that AWS Enterprise Support provides to help you sustainably and cost effectively scale your workloads on AWS.

As part of the Cost Optimization workstream, Technical Account Managers (TAMs) help you create plans to build a FinOps practice across your teams that optimizes spend. The AWS Billing and Cost Management home page provides an overview of your AWS cloud financial management data that comes from AWS Cost Explorer. A recently released alert feature shields you from accidental savings plan procurement and can help you achieve your optimization and sustainability goals.

One of the fundamental offerings for cost optimization is AWS Savings Plans. You can choose from three types of Savings Plans: Compute, EC2 Instance, and SageMaker. This flexible pricing model can help you reduce your bill by up to 72% compared with On-Demand prices, in exchange for a 1-year or 3-year hourly spend commitment for services such as Amazon Elastic Compute Cloud (Amazon EC2), AWS Lambda, and AWS Fargate.

Based on your usage over a selected time period of 7, 30, or 60 days, AWS provides customized Savings Plans recommendations. Before you purchase a Savings Plan, review the recommendations on the AWS Billing and Cost Management console to understand how you can save and the value of the commitment. After you've purchased a Savings Plan, you can't modify, cancel, or resell them.

Solution overview

On January 29, 2024, AWS introduced a new feature that sends an alert when you try to add a Savings Plan to your cart with a commitment amount that's higher than the recommended value.

The preceding figures show alerts that AWS invokes when you try to add items to your cart with a dollar per-hour commitment that exceeds the value that AWS recommends. It's a best practice to adhere to the recommendation calculations so that your Savings Plan operates optimally.

When you compare the average hourly AWS Compute spend in AWS Cost Explorer, it's often higher than the recommended Savings Plans commitment cost per hour. To make the most informed choice, it’s important to understand how AWS determines the recommendations. AWS uses a 7-, 30-, or 60-day baseline to assess your resource usage and recommend a commitment. For example, if you choose 30 days and the AWS recommendation commitment for a Compute Savings Plan is $100 per hour, then your usage for the past 30 days is reviewed and calculated. You then receive a recommendation that spending $100 per hour on AWS Compute services provides you with the larger savings compared with On-Demand rates. The Savings Plans commitment is calculated at an hourly level. Any unused hourly commitment within the time period of the Savings Plan isn't rolled over for future use. To help you monitor your costs, you can use the Cost graph for your Savings Plan.

The following figure shows the AWS Compute Savings Plan recommendation details for Amazon EC2, Lambda, and Fargate cost usage for a period of 30 days. The green shaded area represents the current Compute Savings Plans hourly commitment, and the blue shaded area represents AWS Savings Plans recommended hourly commitment. The pink shaded area represents the estimated On-Demand hourly cost. To avoid losing hours that you've paid for, compare your commitment to your actual spend.

For example, the total usage of AWS Compute services on February 16, 02 PM is approximately $437.42 per hour. The Compute Savings Plan commitment is $200 per hour, so $237.42 per hour of On-Demand usage remains. However, AWS recommends only a $147.528 per hour commitment. If a Compute Savings Plans for $237.42 per hour is purchased, then $89.892 per hour is lost on February 3, 02 PM, and for every hour that's spent that's less than $237.42 per hour. In addition, the average hourly On-Demand spend for the past 30 days is $222.176 per hour. Because the remaining commitment amount doesn't apply to the subsequent hours with higher usage, all the hours where the spend is less than $222.176 are lost if a Compute Savings Plans for average hourly On-Demand spend is purchased.

Although it's a best practice to stay within the recommended guidelines, sometimes you might need to deviate from the AWS Savings Plans recommendations. For example, AWS doesn't consider ad hoc workloads, or workloads that are associated with future releases because they aren't part of past usage. If you expect to run these workloads within the next year, then calculate the minimum hourly usage of the AWS Compute services and then adjust the AWS Savings Plans recommendations commitment dollar per hour value.

When you purchase a Savings Plan, consider these other best practices:

• Purchase in multiple cycles of incremental amounts: You don't always need to purchase the entire AWS Savings Plans hourly commitment up front, especially for AWS Compute services, whose usage frequently shifts. Start with a smaller amount than the recommendation, and then assess your usage and related savings before you make another incremental purchase. You can repeat this process until you've used the available hours in your chosen Savings Plan.

• Right size your compute instances: Before you use Savings Plans, modify any over-provisioned or under-provisioned compute instances so that you get more accurate recommendations.

• Select the best time period for your usage: Choose the number of days (7, 30, or 60) that has the most consistent usage of AWS Compute services so that you buy the optimal amount.

• Subscribe to Savings Plans alerts: Use alerts to notify you before your Savings Plan expires so that you can plan ahead for a new commitment.

Conclusion

Savings Plans are a great cost optimization strategy for your AWS compute workloads. In exchange for a usage commitment, they provide savings beyond On-Demand rates so that you pay only for what you use. If you have an AWS Enterprise Support plan, then you can work with your TAM to determine your usage. AWS Enterprise Support experts can help you assess your workloads by performing an itemized evaluation at each architecture level to arrive at a recommendation. To learn more about how AWS Enterprise Support can help you optimize costs and get the most out of your AWS resources, see AWS Enterprise Support.

About the author

Vivek Appala

Vivek is a Sr. Technical Account Manager at AWS who helps Enterprise customers with cost optimization and successful cloud migration. He has worked with customers from multiple disciplines, such as Travel and Hospitality, Media & Entertainment, Games, and Sports. He helps customers build flexible, scalable, and resilient cloud architectures. In his spare time, Vivek loves composing music and watching movies.