it is a common use case; prior to the bulk implementation, the recommendation was to make parallel requests, which isn't terribly painful under gRPC