AWS’s new Graviton3 EC2 instances are built on its third generation of custom Arm-powered processors. These instances promise up to 25 percent better performance over Graviton2 for compute-intensive workloads. This means that, for applications like distributed data analytics, machine learning, video encoding, gaming, and more, migrating to Graviton3 instances can provide better performance, cost savings, and more energy efficiency.
When you migrate workloads to Graviton3, it’s important to verify that they continue to work properly and perform well. This can help you, for example, rightsize your environment and determine whether you can run the same application on fewer or smaller instances. Datadog provides full visibility into your entire AWS environment, including any Graviton3-powered instances you are running workloads on. This visibility enables you to directly compare performance between architectures and confirm that your application performs well when moving to these new instance types.
Once you enable Datadog’s AWS integration, Datadog will automatically begin collecting telemetry from across all your EC2 instances, including your Graviton3-powered hosts, without any additional configuration. Datadog’s out-of-the-box EC2 dashboard visualizes key CloudWatch metrics from your instances, enabling you to monitor, for example, CPU utilization across each instance type in your environment so you can identify performance issues or improvements in newly provisioned Graviton3 instances.
The Datadog Agent fully supports monitoring both x86 and Arm-based hosts, meaning that once you install it across your instances, you can get even more granular insights into your EC2-hosted workloads. For example, the Host Map provides a birds-eye view of resource usage and other telemetry across your entire environment. You can also use tags to filter, sort, and group hosts to easily monitor system-level CPU or memory utilization across different instance types, clusters, availability zones, or specific applications.
During the migration process, the Host Map makes it easy to validate that your environment continues to perform properly. If you spot hosts that are under significant load, you can then pivot to view all of the processes running on them in real time to determine whether your workloads are performing differently on the new instance types or if there is some other problem.
Having visibility into the performance of similar workloads across different EC2 instance types helps you determine whether your application is showing sufficient performance improvements and continues to operate properly after its migration to Graviton3. Datadog APM enables you to collect and analyze trace data from your services running on different instance types, so you can directly compare request latency and other telemetry. Performing this kind of comparison makes it easier to benchmark application performance and confirm that the new instance types can handle more requests with lower latency.
Datadog has achieved AWS Graviton Ready designation and fully supports monitoring all Graviton-powered workloads no matter where they are hosted. This means that, whether you are deploying applications to provisioned EC2 instances or in serverless environments via AWS Lambda or AWS Fargate, Datadog can give you deep insights into their performance.
Datadog provides full visibility into your Graviton-hosted workloads running in AWS, including those that use the new Graviton3 instance types. By visualizing key resource utilization metrics and application performance data across your instances, you can more easily verify that your workloads have benefitted from migrating to Graviton3. If you’re a Datadog customer, you can enable the AWS integration and deploy the Agent to your instances. If you’re not yet a Datadog customer, you can get started with a 14-day free trial.