By default, wsgi.multiprocess is now False on all threaded WSGI servers. This may be changed by passing the keyword argument multiprocess=True to the servers' constructors.
My original interpretation of wsgi.multiprocess was "is running multiple instances of an application possible with this server implementation?" I think that's a question better answered by the person deploying the application, since even SCGI can be run with load-balancing/redundancy using an HTTP load-balancer.