> > I think this primarily a problem of how web2py is being served rather > than a web2py problem. Just the welcome app and no traffic takes 70MB > of memory by default.
the problem is with the architecture of webserver behind web2py processes vs threaded vs async ok, most of recommended ways of deployment are threaded vs async so what are the options: threaded: servers: apache, nginx, cherokee mode: fcgi, modwsgi, uwsgi (not async) builtin rocket webserver async: any server + uwsgi async unicorn + gevent threaded servers consumes more memory, but async evented servers can't take advantage of multiple processors, a good solution is to spawn a few processes (2 or 4) that will work async like do nginx, unicorn and uwsgi so for webfaction I recommend *uwsgi *with 2 processes async I didn't provide some numbers, but I can say for sure that the difference is very big (a few times less memory)
