On 5 May 2010 20:45, Boris Schaeling <[email protected]> wrote: > On Wed, 05 May 2010 02:49:27 +0200, Graham Dumpleton > <[email protected]> wrote: > >> [...]So, look at database query times and improving them. Look at caching, >> be that in the application using memcached or in front end web server >> caches. Do everything possible to avoid doing the stuff that takes a >> long time and you will get a much better results as far as request >> capacity. > > Thanks for both of your answers! The reason why I've been talking about > processes and threads is that this is the only configuration setting where I > have to guess. And according to the documentation I have to make a good > guess because the number of processes and threads is fixed and doesn't > increase if demand grows. That's why I think it would be great if mod_wsgi > would provide a number to easily measure something like load average > (compare to 'cat /proc/loadavg' to get the load average of a Linux server). > Does it make sense? I wonder because I find it a bit strange that I have to > guess but mod_wsgi doesn't help me to find out if my guess makes sense? Or > am I giving too much weight to this configuration setting (then some > sensible default values in mod_wsgi would make sense so you don't need to > bother at all)?
Problem is that trying to quantify things in a single figure or providing a formula for estimating what to set it to isn't necessarily trivial and ultimately could just be meaningless due to fact that everyones application is different. Did previously do some work on trying to instrument mod_wsgi so as to try and capture some statistics about thread utilisation and other factors, but had to back it out as was all airy fairy and wasn't really obvious how one could even use it to tune things, plus in unfinished state was holding up a release. In the end it just provided more variables to feed the magic voodoo and didn't make it any simpler. This statistics gathering may return at some point when work out what is actually meaningful to capture and how to bring it together in real time where a multiprocess configuration is used. Anyway, if anyone thinks they have a simple solution to the problem, then will gladly listen to suggestions. Graham -- You received this message because you are subscribed to the Google Groups "modwsgi" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/modwsgi?hl=en.
