Sorry about the delay in responding, Google groups never sent me a moderation email or Google is putting its own emails in the spam folder. I am also traveling away from home right now which hasn't help as had limited Internet access.
Can you you post some details about what Apache MPM you are using, what the MPM configuration is, plus whether using mod_wsgi daemon mode and if so what configuration for that. As far as getting much deeper insight into where time is being consumed, suggest start reading about what New Relic is. I work there these days and we may be able to use it to delve into the issue. New Relic has a 2 week pro trial period which will give us access to full features we need to do a deep diving exercise, so don't go signing up immediately and wait until have explored configuration and general issue first so you can get most value from that 2 week pro trial. After the 2 week trial it falls back to Lite level which still provides lots of useful information, but not the slow transaction traces which could be most valuable in sorting out specific bottlenecks in your application etc. For some more details of what you can expect out of New Relic for Django, read: http://blog.newrelic.com/2011/11/08/new-relic-supports-python/ There are also some new abilities in mod_wsgi 4.0 which are working on that may help in working out real problem cases such as blocked requests, but can also explore thread utilization issues as have some experimental stuff in there which reports data that can be injected into New Relic and graphed to see level of request concurrency. Graham On Friday, 13 January 2012, Daniel Benamy <[email protected]> wrote: > First off, thanks for the awesome tool and phenomenal documentation! I > wish more projects had docs even half as good. > > I've been hitting what seem to be pretty restrictive performance > limits with django, mod_wsgi, and apache. This may be an apache > problem, but I'm running out of idea of what to do and who to ask, so > I thought I'd ask here. > > When I run load testing benchmarks, the average response time scales > pretty linearly with the number of concurrent requests I'm making. I'd > expect to see increasing delays if I was making more concurrent > requests than the number of apache workers, apache max clients, or > mod_wsgi daemon threads, but it starts happening with 2 concurrent > requests. > > Here's a graph showing the avg response time for a static file, plain > wsgi hello world app, and django hello world page- > https://docs.google.com/spreadsheet/pub?hl=en_US&hl=en_US&key=0AurdDQB5QBe7dGFncnlUNkdKMVJ4NnYtRjhjaGFIeFE&output=html . > Obviously, if apache is adding some amount of delay for each > additional concurrent request, mod_wsgi and django can't go any faster > than that, but it looks like each layer in the stack is adding some > delay of its own that gets worse with each additional concurrent > request. > > These times are small enough that it's not a big deal if I have 40 > concurrent requests, but the problem is much worse with my full app. > I've got a page that does a whole bunch of work. When I make one > request at a time the response time is on the order of 100 ms. With 10 > concurrent requests, the avg response time is well over a second. All > the work that the page is doing is going to slow things down, but > there's still this linear scaling of response time and I wonder if > it's being caused by whatever is causing the linear performance with > the hello world page. > > Any thoughts or what might be causing this or how to troubleshoot > further? I asked on server fault about the static file part of it and > haven't really gotten anywhere- > http://serverfault.com/questions/344788/why-is-static-page-response-time-going-up-with-increased-concurrent-requests > > Thanks so much, > Dan > > -- > 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. > > -- 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.
