On Mon, Mar 30, 2009 at 9:14 AM, Cosimo Streppone <cos...@streppone.it> wrote: > The problem that we had, this was in November last year, > was that all the backends were at load 15.0-20.0 (normal was ~3-4) > after an update to the application.
That's pretty rare (hitting a CPU problem), and it sounds severe enough that your idea of automated stress tests before deployment would work for catching it. > While I try to get there, I thought it might be useful to dedicate 1 of the > live backends to this "live profiling". Even if the application now is > not having any problem, even at peak times. Sure, give it a shot. My suggestion would be to not collect too much data before analyzing it though, at least at first. Devel::NYTProf can take a really long time if you feed it more than a few requests at once. - Perrin