MSFTCOST07-BP01 Optimize AWS Fargate tasks with AWS Compute Optimizer
AWS Compute Optimizer can be used to help right size and optimize your workloads running on AWS Fargate. If not reviewed, long running tasks can be overprovisioned and increase compute costs.
Desired outcome: Aim to achieve optimal resource allocation and cost efficiency. Right-size our Fargate tasks, avoiding overprovisioning and reducing unnecessary compute costs. Through regular review and implementation of Compute Optimizer's recommendations, we expect to maintain well-optimized workloads that balance performance and cost-effectiveness, ultimately leading to improved resource utilization and significant cost savings in our AWS environment.
Common anti-patterns:
-
Deploying Fargate tasks with initial resource configurations and never reviewing or adjusting them over time, ignoring Compute Optimizer's recommendations and missing significant cost optimization opportunities.
-
Deliberately configuring Fargate tasks with excessive CPU and memory as a precautionary measure, leading to consistent overprovisioning and unnecessary costs despite Compute Optimizer showing opportunities for right-sizing.
Benefits of establishing this best practice:
-
Regular review and implementation of Compute Optimizer recommendations leads to right-sized Fargate tasks, eliminating waste from overprovisioning and reducing monthly compute costs while maintaining required performance levels.
-
Leveraging Compute Optimizer's ML-powered analytics provides objective, metrics-based insights for resource allocation decisions, replacing guesswork with actual usage patterns and ensuring optimal task configurations across your workloads.
Level of risk exposed if this best practice is not established: Medium
Implementation guidance
Enable AWS Compute Optimizer for your organization or account. Establish a monthly review process for Fargate task recommendations. Implement changes gradually, starting with non-production workloads. Monitor performance before and after optimizations. Create a feedback loop to inform future deployments, and consider automating recommendation implementation through Infrastructure as Code practices for consistency and efficiency.
Implementation steps
-
Enable AWS Compute Optimizer in your environment through the AWS Management Console and verify it begins collecting task utilization data for analysis.
-
Schedule monthly review meetings with relevant stakeholders to assess Compute Optimizer's Fargate task recommendations and prioritize which optimizations to implement.
-
Create a change management process that includes testing optimizations in non-production environments first, with clear rollback procedures if needed.
-
Implement approved recommendations through your existing Infrastructure as Code (IaC) framework, ensuring changes are tracked and reproducible.
-
Set up CloudWatch dashboards or alerts to monitor performance metrics post-optimization, ensuring the changes maintain desired service levels while achieving cost savings.
Resources
Related documents:
Related tools:
