A solutions architect wants to cost-optimize and appropriately size Amazon EC2 instances in a single AWS account. The solutions architect wants to…

AWS Compute Optimizer is the designated service for generating right-sizing recommendations for EC2 instances by analyzing CPU, network, and memory metrics. However, EC2 instances do not report memory utilization metrics to Amazon CloudWatch by default. To meet the requirement of optimizing based on memory, the unified CloudWatch agent must first be installed and configured on the instances (Step C). This action enables the collection of memory data. Subsequently, configuring AWS Compute Optimizer (Step D) allows it to ingest this memory data along with the default CPU and network metrics, thereby providing comprehensive and accurate sizing recommendations that fulfill all the requirements.

A. Purchasing a specific support plan is an administrative step. While it may unlock more Trusted Advisor checks, it is not a direct technical action for instance optimization.

B. AWS Trusted Advisor's "Low Utilization Amazon EC2 Instances" check analyzes CPU and network I/O, but it does not analyze memory utilization, thus failing to meet all specified criteria.

E. EC2 Instance Savings Plans are a pricing model for reducing costs on committed usage. This is a financial optimization applied after an instance has been appropriately sized, not a tool for sizing itself.

---

1. AWS Compute Optimizer User Guide: Under the section "Meeting requirements for EC2 instances

" the documentation states

"To get memory utilization recommendations for EC2 instances... The unified CloudWatch agent must be installed on your instances. You must also enable memory utilization metric collection for your instances." This directly links the need for the CloudWatch agent (Option C) to enable Compute Optimizer's (Option D) full capabilities.

Source: AWS Compute Optimizer User Guide

"Meeting requirements for EC2 instances".

2. Amazon CloudWatch User Guide: The section "Metrics collected by the CloudWatch agent" lists the standard metrics available

including memusedpercent. The guide details the process of installing and configuring the agent to publish these custom metrics

which are essential for services like Compute Optimizer.

Source: Amazon CloudWatch User Guide

"Collecting metrics and logs from Amazon EC2 instances and on-premises servers with the CloudWatch agent".

3. AWS Trusted Advisor Documentation: The check description for "Low Utilization Amazon EC2 Instances" specifies the criteria it evaluates: "Checks for instances that were running at any time during the last 14 days and had an average daily CPU utilization of 10% or less and network I/O of 5 MB or less on 4 or more days." Memory is not listed as a criterion.

Source: AWS Support Documentation

"AWS Trusted Advisor check reference"

Cost Optimization category.