Comments on: Profiling puppetmasterd with ruby-prof http://explanatorygap.net/2010/03/11/profiling-puppetmasterd-with-ruby-prof/ many a slip 'twixt mind and lip... Mon, 10 May 2010 22:46:00 +0000 hourly 1 http://wordpress.org/?v=3.3 By: nigel kersten http://explanatorygap.net/2010/03/11/profiling-puppetmasterd-with-ruby-prof/comment-page-1/#comment-2235 nigel kersten Wed, 10 Mar 2010 22:57:09 +0000 http://explanatorygap.net/?p=245#comment-2235 Whoa. That's a much more useful format to analyze in. Thanks Brice! Whoa. That’s a much more useful format to analyze in. Thanks Brice!

]]> By: Brice Figureau http://explanatorygap.net/2010/03/11/profiling-puppetmasterd-with-ruby-prof/comment-page-1/#comment-2234 Brice Figureau Wed, 10 Mar 2010 19:15:56 +0000 http://explanatorygap.net/?p=245#comment-2234 You should try to generate the profile output with the CallTree printer and use KCachegrind to have a better idea of what happens behind the hood. Another idea is to profile only below webrick (ie in the puppet http handler) to profile only the code on which we can do something. And finally, it would be interesting to run the master with JRuby to see if that's actually better or not. Something I want to do since a long time but never really had the incentive to do. You should try to generate the profile output with the CallTree printer and use KCachegrind to have a better idea of what happens behind the hood.

Another idea is to profile only below webrick (ie in the puppet http handler) to profile only the code on which we can do something.

And finally, it would be interesting to run the master with JRuby to see if that’s actually better or not. Something I want to do since a long time but never really had the incentive to do.

]]>