Overview

AWS Compute Optimizer is a web service that provides resource configuration recommendations to help users rightsize their workloads.

This integration enables you to get better EC2 instance type recommendations in AWS Compute Optimizer using memory utilization data from the Datadog Agent. For more information on Compute Optimizer, read What is AWS Compute Optimizer? in the AWS documentation.

Setup

Installation

AWS

1. In the AWS Compute Optimizer console, go to the Accounts page and set your account-level preferences for external metrics ingestion to Datadog.
2. Repeat step #1 for each AWS account you wish to get enhanced recommendations for.

Datadog

3. If you haven’t already, set up the Amazon Web Services integration first for each desired AWS account.
4. Install the Datadog Agent on any EC2 instances to include in the improved recommendations from Compute Optimizer.
5. Install the Datadog - AWS Compute Optimizer integration.

After all steps are completed, it may take up to 30 hours for the recommendations in AWS Compute Optimizer to use the memory utilization data from Datadog.

Validation

Confirm that Datadog is referenced as an External metrics source in the recommendations table for EC2 instances:

How it Works

For all EC2 instances monitored by both Datadog’s AWS Integration and the Datadog Agent, Datadog sends memory utilization data from the Agent to AWS Compute Optimizer to provide enhanced instance recommendations that can potentially lead to cost savings.

Note: The Datadog memory utilization metrics are integrated directly with the AWS Compute Optimizer service and not your AWS account. No additional IAM permissions are needed for this integration since Datadog is not interacting with your AWS account directly.

Data Collected

Metrics

The AWS Compute Optimizer integration does not include any metrics.

Events

The AWS Compute Optimizer integration does not include any events.

Service Checks

The AWS Compute Optimizer integration does not include any service checks.

Troubleshooting

Need help? Contact Datadog support.