Compute Optimizer

Overview

Compute Optimizer is a machine learning service designed to optimize AWS compute resources. It serves two primary purposes:

• Improve cost efficiency by identifying over-provisioned resources

• Enhance performance by rightsizing under-provisioned resources

• Can be implemented at single account or organization-wide level

How It Works

Data Collection Process

1. Service requires access to:

   • Compute resources

   • CloudWatch metrics

2. Tracks performance and utilization patterns over time

3. Analyzes resource usage patterns

4. Produces configuration recommendations

Compatible Resources

1. EC2 Instances

   • Rightsizing recommendations

   • Instance type optimization

2. Auto Scaling Groups

   • Configuration adjustments

   • Scaling parameter optimization

3. EBS Volumes

   • Volume type recommendations

   • Size optimization

4. ECS on Fargate

   • Task size optimization

   • Container size recommendations

5. Lambda Functions

   • CPU allocation optimization

   • Memory allocation recommendations

Key Exam Points

1. Core Functionality

   • Uses machine learning algorithms

   • Analyzes CloudWatch metrics

   • Provides configuration recommendations

2. Service Scope

   • Not limited to EC2

   • Covers multiple compute services

   • Provides comprehensive resource optimization

This service demonstrates AWS's commitment to helping customers optimize their resource usage through automated, intelligent analysis.