View a markdown version of this page

SUS03-BP01 Optimize workloads for high resource usage efficiency - AWS Well-Architected Framework

SUS03-BP01 Optimize workloads for high resource usage efficiency

Optimizing workloads to use resources efficiently and eliminate waste is fundamental to sustainable cloud operations. By implementing a cycle of measurement, optimization, and validation, you can identify and eliminate inefficiencies such as idle resources, over-provisioning, or inefficient processes. The goal is to maximize the business value per resource consumed, reducing both environmental impact and operational costs.

Desired outcome:

  • Workloads operate at high efficiency, using the minimum resources necessary to meet business requirements.

  • You keep resources appropriately sized and actively utilized.

  • You decommission resources when no longer needed, minimizing waste across your cloud environment.

Common anti-patterns:

  • Provisioning resources based on default settings, infrequent peaks, theoretical requirements, excessive SLAs, or unrealistic scenarios rather than actual needs.

  • Failing to monitor resource utilization and identify optimization opportunities.

  • Not taking action to reduce unused or underutilized cloud resources.

  • Not retiring or refactoring inefficient application components.

  • Treating optimization as a one-time project rather than an ongoing process.

  • Optimizing only for cost, and not optimizing usage where there is no cost difference.

  • Limiting optimization to compute only, and not considering other resources or components.

Benefits of establishing this best practice:

  • Lowers the share of hardware-related impacts from raw material extraction to end-of-life treatment (Scope 3 emissions).

  • Lowers the share of required energy for operating infrastructure (Scope 2 emissions).

  • Optimization for resource usage efficiency complements cost- and performance-based approaches. Commercial discounts skew cost-based perspectives. Performance considerations focus on high or peak load scenarios.

  • Contributes to a culture of efficiency and continuous improvement.

Level of risk exposed if this best practice is not established: High

Implementation guidance

Optimizing workloads to reduce waste requires a systematic approach to monitoring, analysis, and improvement. By implementing a continuous optimization cycle, you can identify inefficiencies, implement improvements, and validate their impact on both performance and sustainability.

Implementation steps

  1. Establish baseline metrics for your workload:

  2. Identify optimization opportunities:

    • Use AWS Compute Optimizer to identify right-sizing opportunities and idle resources across compute, storage, databases, and machine learning endpoints.

    • Analyze AWS Cost Explorer, Amazon CloudWatch Metrics, and AWS Trusted Advisor data to identify underutilized resources.

    • Identify services that provide on-demand capacity modes (such as Amazon DynamoDB), auto scaling features (such as Amazon Aurora), or integrate with Application Auto Scaling.

    • Conduct regular reviews of resource utilization and efficiency metrics.

    • Use Application Observability (APM) traces to identify bottlenecks and inefficiencies in distributed applications.

    • Review resource inventory and recent changes.

    • Identify unattached or redundant components.

  3. Implement targeted optimizations:

    • Decommission unused, obsolete, or unattached resources.

    • Refactor inefficient applications, application components, or processes.

    • Right-size over-provisioned resources based on actual utilization patterns.

    • Review auto-scaling policies and performance against observed demand.

  4. Validate improvements and measure impact:

    • Compare pre-optimization and post-optimization RUE metrics to quantify improvements.

    • Verify that performance requirements continue to be met after optimizations.

    • Document lessons learned and successful optimization patterns.

  5. Establish a continuous optimization process:

    • Schedule regular optimization reviews and activities.

    • Automate recurring optimizations where possible, such as removing unattached volumes or upgrading to latest-generation volume types on a schedule with AWS Compute Optimizer automation rules.

    • Build optimization considerations into your development and deployment processes.

Resources

Related best practices:

Related documents:

Related blog posts:

Related videos:

Related examples: