Resource Health shows "Configuration being updated" every night during Azure VM Backup (started recently)

No VM repair or redeployment is indicated: treat this as an informational Azure Backup-correlated control-plane update, then filter UserInitiated Resource Health notifications if they are causing unnecessary alerts.

VirtualMachineConfigurationUpdated means that an authorized user or process updated the VM configuration; Microsoft classifies it as Customer Initiated, Not Applicable, and Informational, with no further action required.^1^ An Available → Available transition also means Resource Health did not detect VM unavailability.

The correlation is strong in this case:

• Scheduled and on-demand backups both reproduce it.
• It occurs immediately after the snapshot.
• changedBy = System.
• The VM returns from Updating to Succeeded.
• Backup jobs succeed.
• Availability remains Available.

Azure Backup starts scheduled jobs by coordinating a point-in-time snapshot. For a running Linux VM using file-system- or application-consistent backup, this involves the VMSnapshotLinux extension.^2^ The VM provisioning state is a control-plane state separate from its power state; Updating means the VM is updating to the latest model and can also cover non-model operations.^3^

The evidence establishes the backup correlation, but it does not document why this only began on 5 October 2026 or why only the B2s VMs emit the annotation. That detail requires Microsoft backend telemetry.

Identify the exact operation

For one reproduced event, compare the backup timestamp with the VM Activity Log:

1. Open Azure portal → Virtual machine → Activity log.
2. Set a narrow time range around Backup now.
3. Filter Resource type to Virtual machines and Resource health.
4. Inspect events such as:
   • Microsoft.Compute/virtualMachines/write
   • Microsoft.Compute/virtualMachines/extensions/write
5. Record the operation name, caller, correlation ID, event timestamp, and status.

The Activity Log records control-plane management operations, and Azure retains these records for 90 days by default.^4^ VM and VM-extension writes are among the Compute operations that can appear there.^5^

Also check VM → Extensions + applications and the Linux extension log:

sudo grep -iE 'enable|snapshot|task|error|warning' \
  /var/log/azure/Microsoft.Azure.RecoveryServices.VMSnapshotLinux/extension.log \
  | tail -n 100

The documented Linux extension log is /var/log/azure/Microsoft.Azure.RecoveryServices.VMSnapshotLinux/extension.log.^6^ Do not uninstall or reinstall the extension while backups succeed; those repair steps apply to extension or backup failures, which are absent here.

Prevent this event from paging

If the alert is intended to detect Azure-caused availability incidents, edit it to notify only for:

• Resource status: Unavailable, Degraded, or optionally Unknown
• Reason type/cause: Platform initiated
• Exclude Available → Available and User initiated

Resource Health alerts support filtering by current/previous status and reason type.^7^ For an ARM-defined alert, add:

{
  "field": "properties.cause",
  "equals": "PlatformInitiated",
  "containsAny": null
}

Microsoft documents this filter specifically to reduce noise from user-initiated events. Be aware that cause can occasionally be null, so a strict filter can omit an early notification whose cause has not yet been populated.^8^

Escalate only if needed

Open an Azure support request under Resource Health – Events and Alerts or Azure Backup Resource Health Insights if:

• the VM becomes Unavailable or Degraded;
• backup starts failing;
• Updating persists instead of returning to Succeeded;
• workload latency or I/O impact is observed; or
• the new behavior since 5 October requires a confirmed service-change explanation.

Include both VM resource IDs, region and SKU, backup job IDs, exact UTC timestamps, correlation IDs, the VirtualMachineConfigurationUpdated annotation payload, and a comparison with an unaffected E4s_v4 VM. Microsoft support can then correlate the Compute and Backup backend operations that are not exposed in customer-visible logs.

References

1. Resource Health virtual machine Health Annotations - Azure Service ...
2. Azure Backup architecture and components
3. States and billing status of Azure Virtual Machines
4. Activity log in Azure Monitor
5. Azure Virtual Machines monitoring data reference
6. VMSnapshot extension
7. Resource Health alerts
8. Create Resource Health alerts with PowerShell and ARM templates