Event Grid queue subscription fails destination validation although the queue exists

Ade Jabar 0 Reputation points
2026-10-05T12:38:48.41+00:00

I am configuring Microsoft Defender for Storage malware-scan results to be delivered through an Event Grid custom topic to an Azure Storage Queue in UK South.

Problem:

Creating the Event Grid event subscription through Azure CLI failed with:

“Endpoint validation: Destination endpoint not found. Resource details: resourceId: <storage-account-resource-id>. Resource should pre-exist before attempting this operation.”

The destination storage account and queue both exist.

Checks completed:

  • StorageV2 account in UK South.
  • Queue existence check using --auth-mode login returns exists: true.
  • ARM GET of the exact queue resource, API version 2023-05-01, returns its ID, name and type successfully.
  • Event Grid custom topic provisioningState is Succeeded.
  • Topic has a system-assigned managed identity.
  • That identity has Storage Queue Data Message Sender assigned at the exact queue scope.
  • Storage publicNetworkAccess is Enabled.
  • Network defaultAction is Allow; bypass is AzureServices.
  • Shared-key access is disabled and must remain disabled.
  • Microsoft.EventGrid and Microsoft.Security providers are Registered.
  • Defender on-upload malware scanning is enabled, with a 10 GB monthly cap.
  • Defender configuration points to this custom scan-results topic.

I then tried a direct ARM PUT using Event Grid API version 2025-02-15 and this body:

{

"properties": {

"deliveryWithResourceIdentity": {

  "identity": {

    "type": "SystemAssigned"

  },

  "destination": {

    "endpointType": "StorageQueue",

    "properties": {

      "resourceId": "/subscriptions/<subscription-id>/resourceGroups/<resource-group>/providers/Microsoft.Storage/storageAccounts/<storage-account>",

      "queueName": "<queue-name>",

      "queueMessageTimeToLiveInSeconds": 604800

    }

  }

},

"filter": {

  "includedEventTypes": [

    "Microsoft.Security.MalwareScanningResult"

  ]

},

"eventDeliverySchema": "EventGridSchema"

}

}

The PUT returned provisioningState: Creating.

Activity Log:

  • Original CLI attempt: Failed at 2026-10-05 11:58:43 UTC, with the endpoint-validation error.
  • Direct ARM attempt: Started then Accepted at 2026-10-05 12:06:17 UTC.
  • The activity-log checks performed afterwards showed no terminal status for the direct ARM attempt.

Subsequent GET requests to the same event-subscription ARM path return:

ResourceNotFound: Event subscription doesn't exist.

Questions:

  1. Why does destination validation report the account missing when both account and queue resolve successfully through ARM?
  2. Is queue-scoped Storage Queue Data Message Sender sufficient for managed-identity destination validation, or is another narrowly scoped permission required?
  3. How can I retrieve the asynchronous operation’s final status when PUT returns Creating but GET returns ResourceNotFound?
  4. Is this configuration supported with shared-key access disabled?

No files have been uploaded to test this route, and successful event delivery has not been demonstrated.

Azure Event Grid
Azure Event Grid

An Azure event routing service designed for high availability, consistent performance, and dynamic scale.


1 answer

Sort by: Most helpful
  1. SHOUMIK CHAKRAVARTY 1,625 Reputation points
    2026-11-14T01:21:52.7933333+00:00

    Hi @Ade Jabar - welcome to the Microsoft Azure QnA.

    Why does destination validation report the account missing when both account and queue resolve successfully through ARM?

    Look at which resource the error names. It reports resourceId: <storage-account-resource-id>, the storage account, not the queue. That is the same scope the documentation requires for the role assignment, which is covered below.

    Is queue-scoped Storage Queue Data Message Sender sufficient for managed-identity destination validation, or is another narrowly scoped permission required?

    The documented requirement is the storage account, not the queue. From Event Grid managed identity delivery:

    “The identity must be a member of the Storage Queue Data Message Sender role on the storage account. It must also be a member of the Storage Blob Data Contributor role on the storage account that you use for dead-lettering.”

    You have it assigned at the exact queue scope. Azure RBAC does not inherit upward, so an assignment at the queue does not grant anything at the account. I would correct that and retest before anything else, since it is the only documented difference between your configuration and Microsoft’s example.

    I could not establish your remaining two questions from the documentation, on retrieving the async operation’s final status and on whether the configuration is supported with shared-key access disabled.

    One separate thing worth a test. The CLI example builds the endpoint as the full queue resource path:

    queueid="$storageid/queueservices/default/queues/<QUEUE NAME>"
    

    and passes it to --delivery-identity-endpoint. Your ARM PUT used resourceId plus queueName, which matches the REST schema and is not wrong, but if you retry through the CLI that is the documented form.

    Help make this community better for everyone: if this answer resolved your issue, please accept it or leave an upvote. If not, share more details in a comment so we can continue the discussion and find the right solution.

    Was this answer helpful?

    1 person found 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.