[BUG] Bump from Azure.Identity 1.5.0 to 1.6.0 throws for ManagedIdentityCredential

Library name and version

Azure.Identity 1.6.0

Describe the bug

After upgrading from Azure.Identity 1.5.0 to Azure.Identity 1.6.0 our applications kept working except for the applications creating the ManagedIdentityCredential throwing the next error from MSI/token/?api-version=2019-08-01&resource=REDACTED&client_id=REDACTED:

ManagedIdentityCredential authentication failed: No Managed Identity found for specified ClientId/ResourceId/PrincipalId.
Status: 400 (Bad Request)

Headers:
Date: Thu, 07 Apr 2022 09:44:52 GMT
Content-Length: 158

See the troubleshooting guide for more information. https://aka.ms/azsdk/net/identity/managedidentitycredential/troubleshoot 

The release note states: ManagedIdentityCredential now attempts to use the newest "2019-08-01" api version for App Service Managed Identity sources. The newer API version will be used if the IDENTITY_ENDPOINT and IDENTITY_HEADER environment variables are set. but our services don’t have the IDENTITY_ENDPOINT and IDENTITY_HEADER environment variables set, at least not where we know about (https://docs.microsoft.com/en-us/azure/app-service/overview-managed-identity?tabs=portal%2Chttp#rest-endpoint-reference) and the “2019-08-01” version is still being used.

Expected behavior

Doesn’t throw when doing the request to MSI/token/?api-version=2019-08-01&resource=REDACTED&client_id=REDACTED for a ManagedIdentityCredential

Actual behavior

Throws when doing the request to MSI/token/?api-version=2019-08-01&resource=REDACTED&client_id=REDACTED for a ManagedIdentityCredential

Reproduction Steps

Bump from Azure.Identity 1.5.0 to Azure.Identity 1.6.0 and ask a ManagedIdentityCredential from MSI/token/?api-version=2019-08-01&resource=REDACTED&client_id=REDACTED

Environment

in an Azure function, if that would matter

[ruben12321]

Not sure what the actual problem was, but when we switched from using the ManagedIdenityCredentials to DefaultAzureCredentials it fixed the issue.

[billpieper]

After additional troubleshooting we have determined that our AKS cluster is impacted by this current outage/issue:

I’m going to wait until the Azure issue is resolved and will respond back here early next week if we are still having an issue or not.