Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Virtual Machine restore points capture the configuration and point-in-time disk snapshots of a virtual machine, enabling granular backup and recovery. You can create restore points at regular intervals to reduce data loss exposure and meet recovery time objectives (RTOs).
For a broader comparison of backup options, see Backup and restore options for Virtual Machines in Azure.
About Virtual Machine restore points
A Virtual Machine restore point stores:
- The Virtual Machine configuration at the time of capture.
- Point-in-time snapshots of all attached managed disks—one disk restore point per disk.
Virtual Machine restore points are grouped into restore point collections, an Azure Resource Manager resource scoped to a specific Virtual Machine. For ARM template examples, see the Virtual-Machine-Restore-Points repository.
Restore points are incremental: the first restore point is a full copy; each subsequent one captures only the changes since the previous snapshot. You can exclude individual disks to reduce storage costs.
Each disk of the Virtual Machine has a corresponding disk restore points within the restore points. For example, if a Virtual Machine has three disks (One OS disk and two data disks) and each restore point has three disk restore points in it.
Once the disk restore points are created, Azure automatically initiates a background data copy from source disk to snapshot.
Consistency modes
| Mode | How it works | Set via |
|---|---|---|
| Application-consistent | Uses Volume Shadow Copy Service (VSS) writers (Windows) or pre/post scripts (Linux) to flush in-flight application data before the snapshot. | Default when consistencyMode is omitted. |
| Crash-consistent | Captures a write-order-consistent snapshot of all disks, equivalent to Virtual Machine state after a power outage or crash. | Set consistencyMode to crashConsistent in the creation request. |
Note: Crash consistency cannot be guaranteed for:
- Disks with read/write host caching enabled (writes during snapshot may not be acknowledged by Azure Storage).
For these configurations, use application-consistent mode.
Restore points for Virtual Machine in scale sets and availability sets
Restore points are created per individual Virtual Machine. To back up all instances in a virtual machine scale set (Flexible orchestration mode) or an availability set, create restore points for each Virtual Machine separately.
Note: Virtual machine scale sets with Uniform orchestration are not supported.
Throttling limits for Restore points
| Scope | Operation | Limit per hour |
|---|---|---|
| Virtual Machine | RestorePoints.RestorePointOperation.PUT (Create new Application Consistent) | 3 |
| Virtual Machine | RestorePoints.RestorePointOperation.PUT (Create new Crash Consistent) | 3 |
| Target restore point collection | RestorePoints.RestorePointOperation.PUT (Copy any Virtual Machine Restore Point) | 3 |
Note
Requests that exceed limits return HTTP 429. Retry after the interval specified in the response.
Limitations
General:
- Restore points are supported only for managed disks.
- Ultra Disks, Premium SSD v2 disks, Write-accelerated disks, Ephemeral OS disks, and shared disks aren't supported for crash consistency mode.
- Ephemeral OS disks, and shared disks aren't supported for application consistency mode.
- A maximum of 500 Virtual Machine restore points can be retained at any time for a Virtual Machine, irrespective of the number of restore point collections or consistency type.
- Concurrent creation of restore points for the same Virtual Machine isn't supported.
- Restore points for virtual machine scale sets in Uniform orchestration mode aren't supported.
- Movement of Virtual Machines between resource groups or subscriptions isn't supported when the Virtual Machine has restore points. Moving the Virtual Machine between resource groups or subscriptions doesn't update the source Virtual Machine reference in the restore point and causes a mismatch of Resource Manager IDs between the actual Virtual Machine and the restore points.
Cross-region copy (Preview):
- Private links aren't supported when copying restore points across regions.
- CMK-encrypted restore points are copied as PMK-encrypted in the target region.
For a complete list of limitations, disk type support, OS support, and API version requirements, see Support matrix for VM restore points.
Troubleshoot Virtual Machine restore points
Most common restore point failures are attributed to the communication with the VM agent and extension. To resolve failures, follow the steps in Troubleshoot restore point failures.
Next steps
- Create a VM restore point.
- Learn more about backup and restore options for Virtual Machines in Azure.
- Learn more about the extensions used with application consistency mode.
- Learn more about how to copy Virtual Machine restore points across regions.