Microsoft.DeviceUpdate RP stuck in 'Registering' state — all instance creates fail with InternalServerError

Kensei Kawamura 80 Reputation points Microsoft Employee
2026-04-30T18:27:03.86+00:00

We are unable to create any new Microsoft.DeviceUpdate/accounts/instances resources in our subscription. Every attempt — via Bicep, ARM, Azure CLI, and the Azure Portal — fails with the same opaque error:

{

"code": "InternalServerError",

"message": "Job failed due to an internal error."

}

Activity log shows no further detail beyond ResourceOperationFailureInternalServerError.

The Microsoft.DeviceUpdate resource provider in this subscription has been stuck in Registering state for 24+ hours.

registrationState : Registering

registrationPolicy : RegistrationRequired

namespace : Microsoft.DeviceUpdate

providerAuthorizationConsentState: (empty)

Forcing re-registration (az provider register -n Microsoft.DeviceUpdate --wait) does not complete and the RP never returns to Registered.

Environment:

  • Subscription: [REDACTED]
  • Tenant: Microsoft TME [REDACTED]
  • Region: West US 2 (account location)
  • API version tried: Both 2022-10-01 and 2023-07-01

What we've verified:

  • Existing ADU resources in the same subscription (created Oct 2025) work normally — only new instance creates fail.
  • The IoT Hub being linked is healthy (provisioningState: Succeeded, state: Active) and has the deviceupdateservice shared access policy.
  • The ADU first-party service principal in this tenant (objectId [REDACTED], appId [REDACTED], displayName Azure Device Update) has been granted IoT Hub Data Contributor on the linked hub.
  • The parent Microsoft.DeviceUpdate/accounts resource provisions successfully and shows provisioningState: Succeeded — only the child accounts/instances resource fails.
  • Resource locks, policies, and quotas are not blocking the operation.
  • All ADU subscription feature flags are NotRegistered, but they were also NotRegistered when our existing working instances were created, so this should not be the cause.

Repro:

  1. In subscription above, create any Microsoft.DeviceUpdate/accounts/instances resource (any name, any RG, link to any healthy IoT Hub).
  2. Resource transitions to provisioningState: Failed within 30s with InternalServerError.

What we need:

  • Confirmation whether Microsoft.DeviceUpdate RP being stuck in Registering is a known issue or service-side problem.
  • Guidance on how to force the RP back to a healthy Registered state, or whether this requires PG intervention.Forcing re-registration (az provider register -n Microsoft.DeviceUpdate --wait) does not complete and the RP never returns to Registered. Environment:
        - **Subscription:** [REDACTED]
      
      
        - **Tenant:** Microsoft TME [REDACTED]
      
        
           - **Region:** West US 2 (account location)
      
           
              - **API version tried:** Both `2022-10-01` and `2023-07-01`
    
    What we've verified:
    • Existing ADU resources in the same subscription (created Oct 2025) work normally — only new instance creates fail.
      • The IoT Hub being linked is healthy (provisioningState: Succeededstate: Active) and has the deviceupdateservice shared access policy.
        • The ADU first-party service principal in this tenant (objectId [REDACTED] , appId[REDACTED], displayName Azure Device Update) has been granted IoT Hub Data Contributor on the linked hub.
                                - The parent `Microsoft.DeviceUpdate/accounts` resource provisions successfully and shows `provisioningState: Succeeded` — only the child `accounts/instances` resource fails.
                  
                    
                       - Resource locks, policies, and quotas are not blocking the operation.
                  
                       
                          - All ADU subscription feature flags are `NotRegistered`, but they were also `NotRegistered` when our existing working instances were created, so this should not be the cause.
                  
                          
                    **Repro:**
          
          1. In subscription above, create any Microsoft.DeviceUpdate/accounts/instances resource (any name, any RG, link to any healthy IoT Hub).
          2. Resource transitions to provisioningState: Failed within 30s with InternalServerError.
      What we need:
    • Confirmation whether Microsoft.DeviceUpdate RP being stuck in Registering is a known issue or service-side problem.
    • Guidance on how to force the RP back to a healthy Registered state, or whether this requires PG intervention.
Azure IoT Hub
Azure IoT Hub

An Azure service that enables bidirectional communication between internet of things (IoT) devices and applications.


1 answer

Sort by: Most helpful
  1. Anshika Varshney 15,030 Reputation points Microsoft External Staff Moderator
    2026-05-01T00:54:25.2+00:00

    Hi Kensei Kawamura •,

    Welcome to Microsoft Q&A .Thank you for reaching out to us.

    The current behavior—where the Microsoft.DeviceUpdate resource provider remains in a Registering state and instance creation consistently fails with InternalServerError—indicates that the provider has not fully completed initialization within the subscription.

    In Azure, a resource provider must be fully registered before any dependent resources can be created. If the provider registration does not complete successfully, provisioning operations may fail without detailed error messages.

    Please check if the following steps help -

    1. Verifying provider registration state -This command helps confirm whether the provider remains in Registering or transitions to Registered.
         az provider show --namespace Microsoft.DeviceUpdate --query registrationState
      
    2. Reviewing Azure Service Health for active advisories
    3. Use the Service Health dashboard to check for any ongoing service issues, planned maintenance, or advisories that may impact resource provisioning.
      1. Azure Service Health provides real-time updates and guidance on service-related incidents affecting subscriptions and regions.
    4. Retrying deployment only after provider becomes Registered Instance creation should be retried once the resource provider successfully transitions to the Registered state, as this is a prerequisite for resource creation.

    The following references might be helpful , please check them out

    Thank you

    Was this answer helpful?

    0 comments No comments

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.