How Do You Reduce Azure Managed Disk Costs?

Azure Managed Disks often account for 15-30% of total compute spend, yet most organizations overprovision storage tiers, leave orphaned disks running, and miss opportunities to decouple performance from capacity. Spotto continuously analyzes managed disk utilization, attachment status, and tier selection across Azure subscriptions, surfacing actionable recommendations to eliminate waste while preserving workload requirements.

The most common savings come from three areas: upgrading to Premium SSD v2 to stop paying for oversized disks, downgrading idle disks attached to deallocated VMs, and deleting orphaned disks that serve no workload. Together, these optimizations typically reduce managed disk costs by 30-50% without degrading performance or availability.

What Problem This Solves

Managed disk cost waste compounds silently because disks bill continuously regardless of VM power state:

Azure Managed Disk Tiers -- Quick Reference

Key distinctions:

How It Works

Unused: Eliminating Orphaned Managed Disks

Are you paying for disks that are not attached to any virtual machine? Orphaned managed disks are unattached to any VM and generate storage charges without serving a workload.

Detection logic:

Detection via Azure Resource Graph:

resources
| where type == "microsoft.compute/disks"
| where properties.diskState == "Unattached"
| extend sku = tostring(sku.name), sizeGB = properties.diskSizeGB
| project name, resourceGroup, subscriptionId, sku, sizeGB, location
| order by sku desc, tolong(sizeGB) desc

Deletion after validation:

# Confirm disk is unattached
az disk show --name diskName --resource-group rg --query diskState

# Create retention snapshot if needed
az snapshot create --source diskId --name snapshotName --resource-group rg

# Delete the orphaned disk
az disk delete --name diskName --resource-group rg --yes

Storage: Right-Sizing and Tier Optimization

Upgrade to Premium SSD v2 (Active Workloads)

Premium SSD v1 forces you to buy more capacity than needed to hit IOPS targets. Premium SSD v2 lets you provision capacity, IOPS, and throughput independently, typically saving 20-40%.

Detection logic:

v2 eligibility checklist:

Migration to Premium SSD v2:

# Review disk utilization over 30 days
az monitor metrics list --resource diskId \
  --metric "Composite Disk Read Operations/sec" \
  --aggregation Average --interval PT1H

# Create snapshot of current disk
az snapshot create --resource-group rg --source diskId --name snapshotName

# Deallocate VM
az vm deallocate --resource-group rg --name vmName

# Create Premium SSD v2 disk with right-sized performance
az disk create --resource-group rg --name newDiskName \
  --sku PremiumV2_LRS --source snapshotId \
  --disk-iops-read-write targetIOPS \
  --disk-mbps-read-write targetMBps --zone zone

# Swap disk and restart
az vm update --resource-group rg --name vmName --os-disk newDiskName
az vm start --resource-group rg --name vmName

Downgrade Idle Premium Disks (Deallocated VMs)

Detection logic:

Downgrade code:

# Confirm VM is deallocated
az vm show --name vmName --resource-group rg --query powerState

# Downgrade disk to Standard SSD
az disk update --name diskName --resource-group rg --sku StandardSSD_LRS

# Tag VM to prevent accidental start
az vm update --name vmName --resource-group rg --set tags.DiskDowngraded=true

# Before restart: restore Premium tier
az disk update --name diskName --resource-group rg --sku Premium_LRS

Automated disk tier management (PowerShell runbook):

# Runbook: Manage-IdleDiskTiers
# Schedule: Daily
Connect-AzAccount -Identity
$vms = Get-AzVM -Status | Where-Object { $_.PowerState -eq "VM deallocated" }
foreach ($vm in $vms) {
    $disks = Get-AzDisk -ResourceGroupName $vm.ResourceGroupName |
        Where-Object {
            $_.ManagedBy -eq $vm.Id -and
            $_.Sku.Name -like "Premium*"
        }
    foreach ($disk in $disks) {
        if ($disk.MaxShares -gt 1 -or
            $disk.Sku.Name -in @("PremiumV2_LRS", "UltraSSD_LRS")) {
            continue
        }
        $tags = @{ OriginalSku = $disk.Sku.Name }
        Update-AzTag -ResourceId $disk.Id -Tag $tags -Operation Merge
        $disk.Sku = [Microsoft.Azure.Management.Compute.Models.DiskSku]::new("StandardSSD_LRS")
        Update-AzDisk -ResourceGroupName $disk.ResourceGroupName \
            -DiskName $disk.Name -Disk $disk
    }
}

Architecture: Zonal Placement and Region Migration

Non-Zonal Placement (Blocks Premium SSD v2)

Migration steps:

Non-AZ Region (Blocks Advanced Storage)

Failure Modes and Mitigations

Cost Impact Summary

Common Questions About Azure Managed Disk Cost Optimization

1. How do I find orphaned disks across all subscriptions?

Use the Azure Resource Graph query provided above. It queries all subscriptions in a single call and returns every disk with diskState: Unattached. For MSPs managing multiple tenants via Azure Lighthouse, the same query works across delegated subscriptions without switching context.

2. Is it safe to delete an unattached disk?

Not always. Before deletion, confirm the disk is not referenced by Azure Backup (Recovery Services vault), Azure Site Recovery, a managed image, or an Azure Compute Gallery image version. If any of these dependencies exist, deleting the disk can break backup restore points or image pipelines. When in doubt, snapshot the disk to Standard storage first and retain the snapshot for your compliance retention period.

3. What is the difference between Premium SSD v1 and Premium SSD v2?

Premium SSD v1 uses fixed tiers where IOPS and throughput are tied to disk size. You must provision a larger disk to get more performance, even if you do not need the capacity. Premium SSD v2 lets you set capacity, IOPS, and throughput independently, so you pay only for what you use. However, v2 is LRS-only, cannot be used as an OS disk, does not support host caching, and requires availability zone placement.

4. Can I downgrade a disk tier while the VM is running?

No. Changing a managed disk's SKU (for example, from Premium_LRS to StandardSSD_LRS) requires the VM to be deallocated. This is by design -- Azure needs to move the disk to different storage infrastructure. Plan tier changes during maintenance windows or automate them using the PowerShell runbook provided above for deallocated VMs.

5. Do disk reservations apply to Premium SSD v2?

No. Azure disk reservations are currently available only for Premium SSD v1 disks, specifically P30 through P80 sizes. Premium SSD v2 and Ultra Disk do not support reservations. If you are migrating from v1 to v2, factor in the loss of reservation discounts when calculating the net savings.

6. What happens to my data when I change a disk's SKU?

Changing the SKU (e.g., Premium_LRS to StandardSSD_LRS) does not erase data. Azure moves the data to the appropriate storage tier in the background. However, the VM must be deallocated during the change, and the operation may take several minutes depending on disk size. Always take a snapshot before any tier change as a precaution.

7. Why can I not use Premium SSD v2 as an OS disk?

Premium SSD v2 does not support host caching, which is commonly used for OS disks to improve boot and read performance. Additionally, v2 disks cannot be used with certain VM features that depend on host-level caching. OS disks should remain on Premium SSD v1 or Standard SSD, while data disks are the primary candidates for v2 migration.

8. How do I automate disk tier management for deallocated VMs?

Deploy the PowerShell runbook provided in this guide as an Azure Automation runbook with a daily schedule. The runbook identifies deallocated VMs with Premium disks, tags the original SKU for restoration, and downgrades the disk to Standard SSD. Before restarting the VM, the reverse operation restores the Premium tier. This approach is fully reversible and typically saves 40-70% on idle disk storage.

9. What is the impact of migrating a VM to an availability zone?

Zonal migration requires planned downtime because you must create a new VM in the target zone. The process involves snapshotting all disks, creating new zonal disks from those snapshots, deploying a new VM with the zonal disks, and updating networking. VMs in availability sets must be fully redeployed. Cross-zone data transfer costs may apply. Estimated effort is approximately 12 hours per VM, but the payoff is access to Premium SSD v2 and Ultra Disk pricing.

10. How does Spotto help with managed disk cost optimization?

Spotto continuously monitors disk attachment status, utilization metrics, tier selection, and zonal placement across all your Azure subscriptions. It automatically flags orphaned disks, identifies Premium SSD v1 disks that would cost less as v2, and detects deallocated VMs with idle premium storage. Rather than running one-off audits, Spotto provides ongoing visibility and alerts when disk waste accumulates -- especially valuable for MSPs managing many tenants.

Use-Case and Scenario Coverage

What to Do Next

Summary