Lumen’s idle disk module classifies idle disks into four distinct categories based on attachment state, VM lifecycle state, mount configuration, and observed I/O activity, enabling precise and context-aware remediation.

Idle Disk Categories

1. Unattached Disks : volumes where the logical connection to a virtual machine no longer exists.

Characteristics:

• Not associated with any active or stopped VM

• Commonly left behind after VM termination, migration, or failed automation workflows

• Continue to generate storage charges despite having no active consumer

2. Reserved Disks : volumes that remain attached to virtual machines that are currently in a stopped or deallocated state.

Characteristics:

• Disk remains logically attached, but the VM is not running

• Often associated with seasonal workloads, paused environments, or forgotten resources

• Charges continue even though the disk is not actively serving workloads

3. Unmounted Disks : volumes attached to running virtual machines but not mapped to any mount point within the operating system.

Characteristics:

• Disk is attached at the infrastructure layer

• No filesystem or mount path exists on the VM

• Frequently caused by incomplete provisioning, failed configuration steps, or legacy volumes

4. No I/O Disks : volumes that are attached to running virtual machines, have valid mount points configured, but have not observed any read or write activity for an extended period.

Characteristics:

• Disk is fully configured and accessible

• No meaningful I/O activity recorded over a configurable time window

• Often associated with deprecated datasets, unused paths, or over-retained storage

Problems addressed:

Lumen’s idle disk module solves critical enterprise pain points below:

• 60-70% of idle disks in enterprises remain undetected (especially reserved, unmounted and no I/O disks)

• Upto 40% of unattached disks are never cleaned up despite detection

• 10-15% of block storage cost is wasted due to dormant disks

Key capabilities:

• Unified dashboard: Cost breakdown by category of idle disks and potential quick wins

• Granular metadata driven insights: Ability to view disk-wise costs and last activity date since which it has been idle

• Disk removal confidence recommendation: High or Low confidence based on historical attachment context, disk age and provisioning origin