Recovery Services vaults that look idle, and why this review raises no finding
What does ZopNight detect here?
Azure Backup bills a Recovery Services vault on two axes, a fee per protected instance priced in 500 GB increments and a charge for backup storage consumed, so a single "idle vault" saving would be guesswork. ZopNight raises nothing from this review and prices vault waste through narrower checks: empty vaults, geo-redundant storage downgrade (not yet priced), and retention tuning.
Signal and threshold
| Field | Value |
|---|---|
| Rule IDs | RC-1379 |
| Category | idle |
| Severity | medium |
| Metric | BackupHealthEvent (context only) |
| Evaluation window | 30d |
| Source | ZopNight |
| Permissions used | Microsoft.RecoveryServices/vaults/read · Microsoft.RecoveryServices/Vaults/backupProtectedItems/read |
Where it applies
Two separate meters on every vault
Azure Backup pricing charges each protected instance a fee that depends on the size of the data backed up, in 500 GB increments, plus the storage the backups consume. Backup storage is billed separately, and on the Standard tier you choose locally redundant, zone-redundant or geo-redundant storage, each at its own rate.
Those two meters move independently. A vault can protect nothing new and still hold years of recovery points, or protect many small VMs with little stored data. Without a way to split the bill between the two, no single number describes what an “idle” vault would save.
Reviewing what a vault still holds
az backup vault list --query "[].{name:name, rg:resourceGroup, location:location}" -o table
az backup item list --resource-group my-rg --vault-name my-vault -o tableProtected items with a stopped protection state, and items for machines that no longer exist, are the ones to question.
What ZopNight looks at for this review
The vault’s backup health events over 30 days are collected as context. Beyond that, this review intentionally does not decide anything on its own: there is no signal here that turns into an action with a price attached.
Why it stays silent
ZopNight only reports a cost finding when the saving is concrete. For a vault, the concrete cases are narrower than “this vault seems idle”, so they have their own checks:
- Azure Recovery Vault Orphan prices deleting a vault that protects nothing.
- Azure Recovery Vault Geo-Redundant Downgrade prices the rate difference of moving backup storage from geo-redundant to locally redundant (not yet priced: it raises no finding until usable per-GB geo-redundant and locally redundant backup rates are available).
- Azure Recovery Vault Retention Tune prices shorter retention on the storage share.
No saving on this page
This review never carries a dollar figure. The checks above do, each for one lever, once they can price it.
Cleaning up a vault by hand
- List protected items and stop protection for machines that were decommissioned, choosing whether to retain or delete their existing recovery points.
- Review backup policies for retention longer than your requirements.
- Consider whether geo-redundant storage is needed for non-production backups.
- Once a vault protects nothing and holds no recovery points, delete it. Azure Backup is moving to soft delete enforced by default (in preview), under which deleted backup data and even a deleted vault pass through a recoverable soft-deleted state first, per Microsoft’s soft delete guide.