I have an issue with my analysis data. My scenario is that we have both old application code and new application code. We expect one of transaction's response time to be high in new application code. But its the otherwise.
Here also the Old application has higher response time than new one which is not expected(since the new code needs tweaking) in our case. So is there a way to find out which code is better with the values of Transactions per second and avg resp time.
Any "upgrade" can make a difference at any time, from subtle to dramatic.
What area(s) of the architecture were impacted by the upgrade? If DB, use a SQL profiler to confirm your timings - though they will read lower than LR timings due to overhead of other architectural components and network.
You could use Ethereal or WireShark to get timings.
Anyway, there is potentially a lot of effort possible with making such a determination. One would also want to know if your scripts execute faithfully from one run to another. Is there any randomization of think time, etc?
Are/were there other processes/tasks involved in the architecture such that they were not running post-upgrade and they were running pre-upgrade? (Anti-virus, batch, sys mgmt. software, etc.)
Also, you have to look at more than just response times and trans/sec.