Hello Roberto, well, about CPU you wrote. I just did one test, and got some interesting results.
I used 2 Apache Bench instances, I started first ONE instance and tried to catch moment, when requests to wsgi never ended, so I had about 10 10-15 seconds to start another apache bench instance, before first one will report about timeouts. so when I started second instance, I saw that all processes of WSGI are loaded at 100% CPU, while first ab instance still waited for last requests results, and 5-10 seconds later just terminated with timeout message. so there are no so called stalled processes. so why then no results for the first apache bench instance? _______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
