Lucas Jans wrote:
I'm seeing a performance issue with 7.3.2. I am currently on 7.1.2 and I thought I would test performance to be sure there are no issues. I have an identical staged server setup. Same hardware. Same database. Same hard drive. Same app server. Only difference is one version was upgraded to 7.3.2. Yet, time-to-first-byte is about 2x worse on 7.3.2, which shocks me considering the focus on performance with 7.3. What could be the cause and how can I diagnose?
A little about our setup - we have a large number of users, portals and tabs in aggregate.
Comparison 1 https://cloudup.com/cjG1_vIwfMd
Comparison 2 https://cloudup.com/cWMGsgthaJJ
Comparison 3 https://cloudup.com/cMqX1uFbnG9
did you do multiple run's of the test and average the scores - it's common for startup of .net app's to vary (which is what time-to-first-byte is predominantly measuring - unless you mean you see the same effect on each page?)
Also, is this just the platform version of one of the evoq solutions?
As to the reasons, it's hard to guess, do you have the ability to use a tool such as redgate profiler , that may lead you to the performance hotspot.
Note: one thing you can try is to set EnableServicesFrameworkTracing to true in web.config (and ensure dotnetnuke.log4net.config is set to All) and see what sort of timings that has, it may reveal something e.g. I noticed the other day that 0.8 seconds of startup time was being consumed by a site where I had multiple dll's that had service framework route registration - so this is clearly an area we can improve. The reason I mention it is that I'm sure there are more service requests in 7.3.1 than 7.1.2, so perhaps that is causing some of the difference (people wouldn't really notice either as most people are worried about page performance and not app start)