loadrunner Curious. In Recent assignments the clients keep requesting browser...

[Raymond Lorenz]

Curious. In Recent assignments the clients keep requesting browser specific performance ie chrome Firefox testing. With http/html I never see a difference in changing the host string or re-recording ( unless the application has specific code, rare). Usually this is closed after explaining we are testing the servers, no browsers are used in replay, and a test to prove it. Question for http/html anyone notice overall differences? Second item I learned a lot of methods to get http scripts to work. But recently people like to jump to truclient at the first issue. Anyone experience this?

 

[James Pulley]

Push that one back to functional testing and development where they can use the built in developer tools to time the inside of the browser client-bound events. Make sure it works for one user/request first, long before it gets to multi-user performance testing. If they wait so long what they will find is that the architectural changes that they need to make to speed up "rendering" will be beyond what time that they have allotted give that they will need to redo a whole functional QA Cycle before they even make it back to performance. I have an upcoming conference presentation coming up on this very topic at the Spring 2016 Software Test Professionals conference about how this obsession with client performance is negatively impacting the ability to assess server performance.