Optimize your AWS costs: 7 concrete levers

Reduce your cloud bill without sacrificing performance

AWS charges based on usage, which means every poorly sized or unused resource increases your bill. Businesses that apply a structured FinOps approach reduce their AWS costs by 20 to 40% on average. Here are seven concrete levers you can activate today.

Lever 1: instance rightsizing

Rightsizing means adapting the size of your EC2 instances to their actual usage. AWS Compute Optimizer analyzes your instance metrics and recommends better-suited types. An m5.xlarge instance running at 15% CPU utilization can be replaced by an m5.large, cutting the cost in half with no performance impact.

This analysis should be performed regularly as needs evolve. A quarterly review of instance sizing is a good practice that generates recurring savings.

Lever 2: Savings Plans

Savings Plans offer discounts of up to 72% compared to on-demand pricing in exchange for a 1 or 3-year consumption commitment. Unlike Reserved Instances, Savings Plans automatically apply to your EC2, Lambda and Fargate instances regardless of region or instance type.

Analyze your consumption over the past 3 months with Cost Explorer to determine your optimal commitment level. Too high a commitment wastes money, too low leaves savings on the table.

Lever 3: S3 storage classes

Amazon S3 offers several storage classes with different pricing. Frequently accessed data stays in S3 Standard. Data accessed less than once a month moves to S3 Infrequent Access (40% cheaper). Archives move to S3 Glacier (up to 95% cheaper).

S3 Intelligent-Tiering automatically moves your objects between storage classes based on access frequency. This service eliminates the need to manually manage transitions and guarantees the optimal rate for each object.

Lever 4: Spot Instances

Spot Instances use unused EC2 capacity at discounts of up to 90%. They are suitable for interrupt-tolerant workloads: batch processing, testing, image rendering, data analysis. AWS can reclaim a Spot Instance with 2 minutes notice.

For critical workloads, combine on-demand instances for the baseline and Spot Instances for peaks. This hybrid approach significantly reduces your average infrastructure cost.

Lever 5: auto-scaling

Auto-scaling automatically adjusts the number of instances based on demand. Your resources increase during traffic peaks and decrease during quiet periods. Without auto-scaling, you pay for maximum capacity at all times.

Configure scaling policies based on CPU utilization, request count or custom metrics. Auto-scaling works with EC2, ECS, DynamoDB and Aurora.

Lever 6: cleaning up unused resources

Forgotten resources are a major source of waste. Unattached EBS volumes, obsolete snapshots, unused Elastic IP addresses and targetless load balancers generate silent costs. AWS Trusted Advisor identifies these resources and recommends their removal.

Set up a monthly review of unused resources. Systematically tag your resources with project and owner to facilitate identification of orphaned resources.

Lever 7: monitoring with Cost Explorer

AWS Cost Explorer visualizes your spending by service, account, tag and period. Configure automatic reports and overspend alerts with AWS Budgets. Visibility into your costs is the prerequisite for any optimization effort.

A FinOps audit identifies quick wins and defines a roadmap for continuous optimization. LCMH performs AWS audits for businesses in Alsace looking to regain control of their cloud spending.

To understand the methodological framework behind these optimizations, read our article on the Well-Architected Framework.

Sources

1. AWS, Cost Optimization Pillar - Well-Architected Framework. docs.aws.amazon.com/wellarchitected/latest/cost-optimization-pillar
2. AWS, Savings Plans. aws.amazon.com/savingsplans
3. AWS, AWS Cost Explorer. aws.amazon.com/aws-cost-management/aws-cost-explorer