Monitoring EBS Capacity Utilization for Better Efficiency

Unlike tracking EC2 CPU usage, AWS doesn’t provide built-in metrics for EBS space usage. To monitor your actual storage usage, you’ll need to collect data at the EC2 instance level. There are a few ways to do this, however they all involve having some sort of agent on the instance itself that will collect this data. The most popular choices to do so include:

1. CloudWatch agent or AWS SSM – These are native AWS agents that allow to collect such metrics through custom CloudWatch metrics or by running OS-level commands where the volume is mounted.
2. APM agents – Such as Datadog or Prometheus, thay collect instance level data, can often collect capacity utilization as well.
3. Datafy’s “read-only” agent – You can use Datafy’s own agent in “read-only” mode which will allow you to better scope the capacity utilization problem.

Once you have this data, the next step is setting up alerts for underutilized storage, prompting timely action. For example, if utilization drops below 20%, you can configure SNS notifications to alert the FinOps or DevOps team (or ideally, the resource owner) to investigate and consider remediation options. While effective, this process can be time-consuming.

Remediation might involve deleting volumes (or other resources). While this is often the most impactful cost-saving strategy, it’s not always feasible. Alternatively, right-sizing volumes may be a solution, though this is more complex. For more on shrinking EBS volumes, you can read Datafy’s blog on  Right-Sizing Your EBS Volumes: Optimizing Storage Without Downtime.