AWS Cost Optimization: 15 Tactics to Cut Your Bill This Quarter | Harness Blog

Answer Capsule: What is AWS cost optimization?

AWS cost optimization is the practice of reducing cloud spend while meeting workload requirements, using tools like AWS Compute Optimizer, Cost Explorer, Budgets, and Cost Anomaly Detection to find waste and act on it. Platforms like the Harness Cost Management Agent extend that native visibility with cross-cloud attribution and automated remediation.

Quick Facts

AWS bills can grow quickly as teams add workloads, scale infrastructure, and adopt new services. The FinOps Foundation's State of FinOps 2026 report, based on responses from 1,192 respondents representing more than $83 billion in annual cloud spend, found that workload optimization remains the single largest current priority for FinOps practitioners, even as the discipline's scope expands into AI, SaaS, and other technology spend. For AWS teams, that means reviewing resource utilization, pricing commitments, storage, networking, and cost visibility to identify opportunities to reduce spend without affecting workload requirements.

Here are 15 tactics you can evaluate and implement to reduce AWS costs without compromising performance or reliability.

Why are AWS bills so hard to optimize?

AWS makes it easy to provision resources, but that flexibility can make costs harder to track and control as environments grow. Across multiple teams and accounts, small inefficiencies can accumulate into significant unnecessary spend. A development team might leave an EC2 instance running outside working hours, while a production workload may be sized for occasional peak demand even though utilization is lower most of the time. 

Teams need to understand what drives spend, who owns it, and whether a proposed change could affect workload. A monthly AWS bill shows what was spent, but not necessarily why costs changed or where to act. Cost allocation, utilization data, and actionable recommendations give teams the context to prioritize optimization opportunities.

Once you know what is driving spend, prioritize the opportunities with the greatest potential impact.  Not every optimization requires an architectural overhaul or a long-term commitment. Some changes, such as removing unused resources or adjusting instance sizes, can quickly reduce spend. Others, such as AWS Savings Plans, storage tiering, and network optimization, require a closer look at usage patterns and workload requirements. The following 15 tactics cover the main opportunities across compute, storage, networking, and cost visibility, focusing on changes teams can evaluate and implement this quarter.

How can you reduce AWS costs with these 15 tactics?

To reduce AWS costs effectively, prioritize tactics based on where your AWS environment has the most waste and how quickly you can act. Idle resources and rightsizing opportunities can often be addressed without architectural changes, while pricing commitments, storage transitions, and network changes require more analysis. The tactics below cover compute, storage, networking, and cost visibility so teams can target savings without compromising workload requirements.

What are the most effective compute cost tactics?

Compute is a major source of AWS spend, and unused or overprovisioned capacity can create avoidable costs. Start by rightsizing before committing to discounted capacity. AWS's 2026 analysis of more than 71,000 opted-in customers found that larger customers combining AWS Savings Plans with active rightsizing improved their median Cost Efficiency score four times faster than customers using Savings Plans alone. AWS also found that only 47.1% of customers with Savings Plans coverage were actively taking rightsizing actions.

1. Right-size EC2 instances. Compare actual CPU and memory utilization with the instance size you're paying for, and use AWS Compute Optimizer to identify suitable alternatives. This is not simply about choosing a smaller instance: workload performance and capacity requirements still need to be validated. In one AWS customer example, GE Vernova saved $504,000 by downsizing 50 instances.

2. Use Savings Plans for predictable compute usage. If a portion of your compute demand is steady, a Savings Plan can lower the effective rate in exchange for a commitment. Estimate your baseline after rightsizing, then use that baseline to determine the appropriate commitment. AWS's analysis indicates that combining pricing commitments with ongoing rightsizing is more effective than relying on Savings Plans alone.

3. Use Reserved Instances selectively. Reserved Instances can make sense for workloads with stable, predictable usage and configuration requirements. For EC2, compare RIs with Savings Plans based on your workload's flexibility needs before committing. AWS offers discounts of up to 72% for eligible Standard Reserved Instances, depending on the configuration and term.

4. Use Spot Instances for interruption-tolerant workloads. Spot Instances use spare AWS capacity at a lower price than On-Demand, but AWS can reclaim that capacity when it needs it. They work well for batch processing, fault-tolerant distributed workloads, and development environments that can recover from interruptions. AWS offers Spot discounts of up to 90% versus On-Demand, while Amaysim reports 75% lower compute costs after adopting Spot across its workloads.

5. Schedule non-production resources. Development and test environments may not need to run outside working hours. Automating start and stop schedules can reduce compute and database costs during periods when those environments are not in use. Before automating shutdowns, account for application dependencies, persistent data, and workloads that require continuous availability. AWS documents savings of up to 70% for scheduled non-production environments.

How can you optimize storage and networking costs?

Storage and networking costs often grow as workloads scale, making them easy to miss during resource-level optimization. Match storage costs to data access patterns, automate retention and tiering, and review network paths for unnecessary data processing and cross-Availability Zone charges.

6. Match S3 storage classes to access patterns. Choose storage classes based on how frequently data is accessed and how quickly it needs to be retrieved. S3 Intelligent-Tiering is useful when access patterns are unpredictable because it automatically moves eligible objects between access tiers as their access patterns change, without requiring you to manually manage those transitions.

7. Apply S3 Lifecycle policies. Use lifecycle rules to automatically transition objects to lower-cost storage classes or expire them when you no longer need to retain them. This is useful for data such as logs, backups, and archives with defined retention requirements. Check transition and retrieval costs before applying a policy, particularly when objects may still be accessed frequently.

8. Remove unattached EBS volumes. Terminating an EC2 instance does not necessarily delete its associated EBS volumes, which can continue to incur storage charges. Review unattached volumes regularly and remove them only after confirming they are not needed for recovery, backups, or other workloads.

9. Move eligible EBS volumes from gp2 to gp3. Where workload requirements allow, migrating from gp2 to gp3 can lower storage costs while providing more flexible control over IOPS and throughput. AWS says gp3 can cost up to 20% less per GB than gp2. Review the volume's current utilization and performance requirements before migrating, rather than treating the lower storage price as the only consideration.

10. Review NAT Gateway and data-transfer paths. NAT Gateways incur hourly and per-GB data-processing charges. For private workloads that access supported AWS services such as S3 or DynamoDB, gateway VPC endpoints can keep that traffic off the NAT Gateway and avoid its processing charges. Also review cross-Availability Zone traffic and network placement for high-volume workloads to identify unnecessary transfer costs.

How can you improve AWS cost monitoring and visibility?

Cost visibility is a core part of AWS cost management, turning individual optimization efforts into an ongoing process. AWS cost optimization tools can help teams understand how spending changes, connect costs to the workloads and teams responsible, and identify unusual increases early. AWS provides capabilities for each of these tasks, from cost analysis and allocation to budgets and anomaly detection.

11.  Use AWS Cost Explorer to identify cost drivers. Break spending down by dimensions such as service, account, Region, usage type, and time period to isolate the sources of cost changes. Use these views to identify which services or workloads warrant deeper investigation before making optimization changes.

12.  Use AWS Data Exports with CUR 2.0 for deeper analysis. When Cost Explorer doesn't provide enough detail, AWS Data Exports can deliver granular cost and usage data through the Cost and Usage Report 2.0 schema. Teams can query this data to investigate costs across large or complex environments and build customized reporting beyond the views available in Cost Explorer.

13.  Allocate costs to teams, applications, and environments. Use cost-allocation tags and Cost Categories to associate AWS spending with business or engineering dimensions such as team, application, environment, or cost center. Consistent allocation makes it easier to assign ownership, compare spending across workloads, and investigate unexpected changes.

14.  Set budgets and spending alerts. Create budgets for accounts, services, workloads, or cost categories and alert owners when actual or forecast spending approaches a defined threshold. Use these alerts to trigger an investigation before spending moves significantly beyond expectations.

15. Enable Cost Anomaly Detection. Configure anomaly monitors to flag unusual spending patterns and route alerts to the teams responsible for investigating them. AWS Cost Anomaly Detection supports monitoring across dimensions such as linked accounts, cost allocation tags, and Cost Categories, helping teams trace unexpected changes to the relevant spend area.

Predictable compute usage also creates an opportunity to reduce rates through AWS commitment-based pricing, an important part of AWS cost management for steady workloads. Savings Plans and Reserved Instances offer different levels of flexibility and apply differently across AWS services, so the choice should reflect your workload's stability and configuration requirements.

Savings Plans vs. Reserved Instances: Which saves more?

Both can reduce compute costs compared with On-Demand pricing, but they differ in how much flexibility they provide. Savings Plans generally offer more flexibility for EC2 workloads that may change instance types, Regions, or other configuration characteristics. Compute Savings Plans can apply across EC2 instance families, sizes, Regions, operating systems, and tenancies, as well as Fargate and Lambda. EC2 Instance Savings Plans are more targeted and apply to a specific instance family in a Region.

Choose Savings Plans when compute demand is predictable but infrastructure may change over time. Consider Reserved Instances when a workload has a stable configuration and capacity requirements. In either case, right-size first and base the commitment on the usage level you expect to sustain.

How much can AWS cost optimization save?

Savings potential varies by workload, utilization, and configuration. The figures apply to the affected resource or cost area, not the total AWS bill. “Indirect” refers to tactics that improve visibility and control rather than directly reducing spend. The ranges below are directional estimates based on AWS pricing and optimization guidance and customer examples cited in this article. Actual savings depend on workload, utilization, configuration, and the amount of avoidable spend. 

* Up to 72% applies to EC2 Instance Savings Plans and eligible Reserved Instances. Compute Savings Plans offer discounts of up to 66%. Actual savings depend on the commitment, workload, and pricing configuration.

Source note: AWS savings estimates are based on specific workload and configuration assumptions. Actual results will vary.

These tactics can reduce specific sources of AWS waste, but keeping savings on track across a growing environment requires continuous visibility, optimization, and governance. This is where Harness Cost Management Agent can complement AWS's native cost-management capabilities.

How Harness supports AWS cost optimization

AWS provides a strong foundation for monitoring and managing cloud costs, but teams with complex environments may need to connect cost visibility with optimization and action. Harness approaches this challenge by bringing those capabilities together so teams can identify waste, act on recommendations, and maintain cost controls over time.

The challenge

AWS-native tools provide the data and controls needed to manage cloud costs. As environments grow across accounts, teams, and workloads, however, identifying an opportunity is only part of the problem. Teams also need to connect cost data with the resources and owners responsible for taking action, while ensuring optimization changes do not create operational issues.

This can make cost optimization an ongoing operational task rather than a one-time cleanup exercise. Teams need visibility into where spending is changing, recommendations that help prioritize opportunities, and controls that can support action at the resource level. 

The approach

The Harness Cost Management Agent attributes, optimizes, and governs cloud and AI spend at the level of autonomy you choose — Recommend, Approve, or Autonomous. This helps teams move from finding waste to acting on it and preventing it from returning.

The agent connects cost visibility with engineering action through granular cost insights, automated optimization, and policy-based governance. Its capabilities include rightsizing recommendations, idle-resource management with AutoStopping, autoscaling, automated commitment management, and governance controls that can act on your behalf rather than just flag issues for review.

The outcome

Connecting these capabilities helps teams move beyond identifying individual sources of waste and build a more continuous approach to cloud cost optimization. They can prioritize opportunities, automate appropriate remediation, and apply governance as cloud usage changes.

Harness Cost Management Agent helps teams move from one-time cost cleanup to continuous cloud cost optimization. 

How are Harness customers reducing AWS cloud costs?

Customer results show how these capabilities translate into measurable cloud cost savings. They also demonstrate how visibility and automated optimization can address different sources of cloud waste.  

How did OneAdvanced achieve 33% annualized cloud cost savings? 

OneAdvanced was dealing with budget overruns and limited adoption of its previous cloud cost management tool. After implementing Harness's cloud cost management platform (now the Harness Cost Management Agent), the team gained visibility across its AWS accounts, used anomaly detection to flag potential overspend, and applied recommendations and AutoStopping to reduce waste.

The company has achieved 33% annualized cloud cost savings to date.

“At OneAdvanced, we felt like we had a good handle on cloud cost governance. Then, we implemented Harness Cloud Cost Management. And what we actually discovered was that we were just okay. Harness CCM took us to excellence, and it has made huge savings for us!”

Jay Patel, Director of Platform Engineering-DevOps, OneAdvanced  

Source: OneAdvanced cuts cloud costs by 33% with Harness CCM 

How did Tyler Technologies achieve $1.2 million in annualized savings?

Tyler Technologies had thousands of non-production environments with different usage patterns, making fixed shutdown schedules ineffective. The team reorganized its cloud estate around client time zones and activity patterns and used Harness Cloud AutoStopping to shut down idle environments automatically.

The approach produced $1.2 million in annualized savings after six months. Tyler initially saw $15,000 to $20,000 in monthly savings, eventually reaching more than $100,000 in monthly savings.

“Right away we were saving $15,000 to $20,000 a month, and in only six months, we saw our first savings topping $100,000 a month.”

Chris Camire, Senior Manager of Technical Services, Tyler Technologies

Source: Tyler Technologies Saves $1.2M Annually with Harness CCM 

What should you prioritize first?

The customer examples show that cost optimization can come from different parts of an AWS environment, from idle resources to inefficient resource configurations. The right starting point depends on where your environment has the greatest opportunity for improvement.

Start with the areas where your environment shows the clearest opportunity, such as idle resources, overprovisioned compute, inefficient storage, or unexpected spending. Then use cost visibility, ownership, and monitoring to track whether those changes continue to deliver value as workloads change.

The goal is not simply to reduce the next AWS bill. It is to build a repeatable approach to managing cloud spend while maintaining the performance, availability, and reliability your workloads require.