Share via

Unable to delete account from Azure Managed Redis

Prashanth Talari 0 Reputation points
2026-04-15T15:56:13.29+00:00

We are experiencing an issue with Azure Managed Redis (Redis Enterprise) where user‑assigned managed identities (UAMI) added as Redis users cannot always be deleted.

Observed behavior:

Redis users are created using Bicep (Microsoft.Cache/redisEnterprise/databases/accessPolicyAssignments)

Deleting a Redis user from the Azure Portal shows “Deletion may take a few minutes” but never completes

No error message is returned via Portal, Azure CLI, ARM, or REST API

The Redis user remains indefinitely, and some entries display only GUIDs instead of friendly identity names

There is no supported force‑delete option for Redis Enterprise users

Steps tried to resolve:

Retried deletion multiple times over several days

Verified managed identity still exists in Entra ID

Re‑deployed access policy assignments using Bicep

Checked Azure CLI support for Redis Enterprise user deletion

The only way found to clear the stuck Redis users was to delete and recreate the Redis Enterprise instance, which is not a viable solution for production environments.

Questions:

Is this a known issue or bug in Azure Managed Redis (Redis Enterprise)?

Is there a supported way to fully delete stuck Redis users without recreating the Redis instance?

Is there a planned fix or recommended best practice when using managed identities with Redis Enterprise?

References used:

Bicep template:

https://learn.microsoft.com/en-us/azure/templates/microsoft.cache/redisenterprise/databases/accesspolicyassignments?pivots=deployment-language-bicep

CLI reference:

https://learn.microsoft.com/en-us/cli/azure/redisenterprise/database/access-policy-assignment

Thanks,

Azure Cache for Redis
Azure Cache for Redis

An Azure service that provides access to a secure, dedicated Redis cache, managed by Microsoft.


2 answers

Sort by: Most helpful
  1. Saraswathi Devadula 16,015 Reputation points Microsoft External Staff Moderator
    2026-04-16T04:28:45.4633333+00:00

    Hello Prashanth Talari

    It looks like the “stuck” Redis users you’re seeing are coming from the Data Access Policy feature in Azure Managed Redis Enterprise, which is still in preview and currently doesn’t support clean deletion of user-assigned identities once they’re added. At this time:

    Recommended next steps:

    1. If you absolutely need per-identity access control today, consider using a non-Enterprise Azure Cache for Redis with Azure RBAC (AAD) instead of the Preview Data Access Policy.
    2. Bicep template for accessPolicyAssignments: https://learn.microsoft.com/azure/templates/microsoft.cache/redisenterprise/databases/accesspolicyassignments?pivots=deployment-language-bicep
      Azure CLI access-policy-assignment ref: https://learn.microsoft.com/cli/azure/redisenterprise/database/access-policy-assignment?view=azure-cli-latest
    3. Could you please share the details requested in the private message for further troubleshooting.
    0 comments No comments

  2. Q&A Assist
    2026-04-15T15:56:46.9+00:00
    AI generated content. This question contains an answer created with AI from Q&A Assist. Learn more

    This behavior matches known intermittent issues with managed identity assignment and removal in Azure resources, but the context does not document a Redis Enterprise–specific bug or a supported force‑delete mechanism for stuck Redis users.

    From the available information:

    1. Known issue pattern with managed identities
      Managed identities can occasionally get into a bad assignment state where operations such as assign/unassign fail silently or with confusing errors. The documented guidance for such cases is:
      • For user‑assigned managed identities: remove the identity assignment from the resource and then reassign it.
      • For system‑assigned identities: disable the identity and enable it again.
      This pattern is described as a general managed identity issue, not specific to Redis, but the symptoms (resource thinks an identity is still attached, operations blocked or stuck) are similar to what is being observed with Redis Enterprise users.
    2. No documented force‑delete for Redis Enterprise identities
      The context only documents:
      • How to create and manage Azure Managed Redis / Redis Enterprise resources via az redisenterprise.
      • How to remove managed identities from Azure Cache for Redis using az redis identity remove (for the non‑Enterprise Microsoft.Cache/Redis resource).
      There is no documented “force delete” or low‑level cleanup command for stuck Redis Enterprise access policy assignments or users. The only destructive cleanup shown is deleting the cache or resource group itself.
    3. What is documented as supported remediation
      Based on the general managed identity guidance, the only supported remediation steps that align with the documented best practices are:
      • For user‑assigned managed identities used as Redis users:
        1. Remove the managed identity assignment from the Redis resource (or database/access policy assignment) using the supported management plane (Bicep/ARM/CLI/Portal).
        2. Reassign the user‑assigned managed identity to the resource.
      • If the identity itself is no longer needed anywhere, delete the user‑assigned managed identity resource and then manually clean up any role assignments that reference it.
      For system‑assigned identities (if used anywhere in the flow):
      • Disable the system‑assigned identity on the resource and then enable it again.
      These steps are the only documented workaround for identity assignment issues; there is no additional Redis‑specific cleanup guidance in the context.
    4. Best‑practice considerations
      The documented best practices for managed identities emphasize:
      • Prefer managed identities over keys for authorization where possible.
      • Carefully manage lifecycle: user‑assigned identities outlive resources and must be explicitly deleted; role assignments tied to deleted identities are not automatically removed and must be cleaned up.
      Applied to Redis Enterprise:
      • Design access policy assignments so that user‑assigned identities are reused and not frequently created/deleted per deployment, reducing the chance of hitting rare assignment bugs.
      • Periodically audit role assignments and identity usage, and manually remove role assignments that reference deleted or unknown identities.
    5. Answers to the specific questions
      • Is this a known issue or bug in Azure Managed Redis (Redis Enterprise)?
        The context only documents a general “rare” issue with managed identity assignment operations across Azure resources, not a Redis Enterprise–specific bug.
      • Is there a supported way to fully delete stuck Redis users without recreating the Redis instance?
        The context does not document a Redis Enterprise–specific force‑delete or cleanup mechanism. The only documented remediation for identity‑related stuck states is to remove and reassign managed identities (or disable/enable system‑assigned identities) and to manually clean up role assignments.
      • Is there a planned fix or recommended best practice when using managed identities with Redis Enterprise?
        No planned fix is documented in the context. The only relevant best practices are the general managed identity recommendations: manage lifecycle carefully, delete unused user‑assigned identities, and manually remove stale role assignments.

    References:

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.