@bentley , we’re encountering the same issue. Is this a bug? Or are we doing something “wrong” in our implementation?