An Azure service that is used to provision Windows and Linux virtual machines.
Azure Support’s security review is the required reactivation path; update the existing request with the production impact and request the highest severity available under your support plan rather than attempting billing or self-service reactivation.
Escalate the existing request
- In the Azure portal, open Help + support → All support requests, then select the existing request.
- Select New message and explicitly request:
- Immediate review by the team handling suspicious-activity/AUP restrictions.
- Reactivation of the subscription.
- Assistance preserving or exporting the data if reactivation cannot be completed promptly.
- Confirmation of any deletion deadline stated in the notification.
- Include a concise incident summary:
- Exact disablement time and current outage duration.
- Production customer impact.
- Subscription type: Pay-As-You-Go.
- The
ReadOnlyDisabledSubscriptionerrors. - The simultaneous VM stop timestamps and
VMStoppedToWarnSubscriptionstatus. - Confirmation that the balance is zero and that the Activity Log contains no user or service-principal stop operation.
- A direct request for an estimated next update time.
- Select Change severity:
- If no engineer is assigned, select the highest available severity and choose Change.
- If an engineer is assigned, send a message requesting the severity change.
- If the assigned engineer is unavailable and the change is urgent, Microsoft documents contacting customer service and asking the agent to change the severity. The maximum severity available depends on the support plan.^2^
Managing the request requires Owner, Contributor, or Support Request Contributor at subscription scope. The same request page also supports uploading relevant evidence, such as exported Activity Log events, screenshots, timestamps, and the notification email.^2^
Because this is neither an overdue-balance case nor an administrator-canceled subscription, the portal’s ordinary Reactivate workflow and payment remediation do not apply. A production outage also does not itself bypass the security review.
Data export while disabled
The disabled state explains why the VMs cannot be started or changed: Azure blocks resource creation and updates (PUT/PATCH) and action operations (POST). Resources can still be retrieved or deleted (GET/DELETE), storage becomes read-only, and VMs are deallocated.^3^
This distinction limits what can be exported:
- Existing blob data might remain readable if valid data-plane credentials or a previously issued SAS URL already exist.
- Creating a new managed-disk export URL is not merely a read. It requires
az disk grant-access, which generates a SAS for the managed disk.^4^ - Because the subscription rejects action operations, generating that new disk SAS is not a viable self-service export path in the current state.
Add this explicit request to the support case:
If immediate reactivation is not possible, please preserve all resources and provide an approved method to export the managed disks and storage data, or temporarily restore the operations required to generate read-only export access. Please also confirm the applicable data-retention/deletion deadline.
Do not delete resources as an attempted workaround; deletion is destructive and does not create an export.
If an existing Blob Storage SAS or other working data-plane authorization is available, download accessible blobs immediately with AzCopy. For example:^5^
azcopy copy `
'https://<account>.blob.core.windows.net/<container>?<existing-SAS>' `
'D:\AzureRecovery\<container>' `
--recursive
This does not help with managed disks unless a usable disk SAS was generated before disablement.
After reactivation
Most resources resume automatically, but Microsoft recommends checking each service and restarting resources where necessary. If resource management remains unavailable for more than 30 minutes after reactivation, contact Azure Billing Support through the support case.^6^
For the VMs, verify:
- The subscription status is Enabled.
- VM, disk, NIC, and public IP resources still exist.
- Required IP addresses have not changed—temporary IP addresses can be released while a subscription is disabled.^3^
- Each VM starts successfully.
- Application dependencies, queues, databases, DNS, and customer-facing endpoints pass health checks.
References