Skip to content

AWS / Technical guide

Review EBS spending without losing required data

An unattached EBS volume can still hold the only usable copy of application data. An attached volume can also have excessive capacity or provisioned performance. These conditions require different checks and should not share one delete recommendation.

Start with the volume's purpose, attachment history and recovery policy. A current Available state identifies an unattached volume; it does not prove that the data is disposable.

Discuss EBS review coverage

Separate unattached storage from overprovisioned storage

For an unattached volume, identify the previous workload, owner and required retention period. Determine whether a snapshot exists, whether it is restorable and who can access the encryption key. A snapshot that nobody can restore is not a sufficient recovery plan.

For attached storage, compare capacity and provisioned IOPS/throughput with actual workload demand and the instance's EBS limits. A faster volume cannot overcome a lower instance ceiling. Changing type or performance requires validation of the application's latency and recovery expectations.

EBS review paths
ConditionReview before action
Unattached volumeData owner, retention, restore and planned reuse
Low capacity useGrowth, filesystem and migration constraints
Low performance useObserved peaks, instance ceiling and latency needs
Old snapshotRecovery chain, retention and restore usability

A recovery copy has its own cost and controls

AWS documents taking a snapshot before deleting an unneeded volume and notes that deletion removes the volume's data. Recycle Bin retention rules can change whether deletion is immediate. Review the actual retention configuration instead of assuming the meter or recovery obligation ended at the delete request.

Keep the replacement storage and retention cost in the proposed scenario. Do not count the full original volume charge as invoice savings if another retained resource or commitment remains. Snapshot deletion also needs application-specific recovery review.

Validate performance changes separately from deletion

For a volume-type or provisioned-performance change, use AWS-supported modification or migration methods and document the constraints. Do not assume allocated capacity can be reduced in place. A smaller destination may require a data migration and validation process.

Capture latency, throughput and error behavior before and after the change. Define rollback around the service, not merely whether the modification request succeeded. Compare the affected charge components over a suitable billing window.

  • Review the workload's peak and maintenance behavior.
  • Preserve encryption and recovery access.
  • Include snapshot and migration costs in planning.
  • Obtain the data owner's approval before any destructive action.

Costframe's current EBS boundary

Costframe's AWS pilot includes an unattached-volume review path based on supported inventory and cost evidence. It does not inspect application data or establish that every volume is safe to delete. Unknown costs and capability gaps remain limitations.

Do not assume a complete EBS performance optimizer or snapshot-chain analysis is available from that finding. Costframe does not issue customer-cloud deletion or modification calls. The approved change belongs to your team.

References and further reading

Provider documentation explains native capabilities. Costframe availability and limitations are stated separately in this guide.

Confirm the storage evidence in the pilot

Ask which volume states and billing records Costframe can review for your intended AWS scope.

Discuss EBS review coverage