Do interceptor failures fall under transient fault handling?

Hi,

Currently seeing an issue where failure in auth interceptor (http exception calling ADFS) is propagated to caller as Error starting grpc call, StatusCode=Unavailable. That is fine, except that Retry Policy for given service is to retry Unavailable calls 3 times. Is that a misconfiguration on my part somewhere, or do interceptors do not fall under the policy, and retries inside of interceptor should be handled in the interceptor?

Do you need a sample repro?

Start with the linked gRPC documentation and troubleshooting pages, then investigate how the service retry policy treats failures from authentication interceptors. A useful result would be a minimal reproduction or a documented determination of whether interceptor failures are retried and where such failures should be handled.

Written by the indexing model from the issue text.