Pravesh - when you ask about "branch marks", are you referring to techniques for structuring your tests so metrics are collected & aggregated at a meaningful level?
*My* technique is to place all the requests needed to complete a single user-facing action into their own Transaction Controller. So, when my script performs "Display the Home Page", "authenticate the user", "generate the TPS report" or "delete all items from shopping cart", those are each their own Transaction. JMeter's Summary Report listener, as well as Taurus and BlazeMeter will total & display the metrics for all the requests (and optionally include Think Time) within each Transaction Controller. Another technique I use with Transaction Controllers is to ONLY select "Include duration of timer and pre-post processors in generated sample" in situations where a *user is waiting* for the system to complete some processing action. For example, we often measure how long the system takes to complete a file upload. We insert 30-second Timers in a loop that checks the upload status. By selecting "Include duration of timer..." the metrics for that transaction more accurately reflect the total upload time. In situations where the *system is waiting* for the user (e.g., typing in username/password), we de-select "Include duration..." so think times don't skew our metrics. Steve Kram -----Original Message----- From: Dmitri T <glin...@live.com> Sent: Wednesday, August 2, 2023 2:48 AM To: JMeter Users List <user@jmeter.apache.org>; Pravesh <prajapati.prav...@gmail.com>; user-h...@jmeter.apache.org Subject: Re: Performance testing queries Pravesh wrote: > Hi Team > Have some queries related to what values should we use for below in > Jmeter > > 1. How we should calculate think time? > I mean what value we should use. > 2.what are the major matrix should we measure? > 3.what branch mark should we follow and what is standard ? > 4.if we are doing performance testing for component, should we use > think time or not? > > Regards > Pravesh p 1. The main idea is to increase the load gradually, if you don't have better suggestions you can stick to JMeter documentation <https://jmeter.apache.org/usermanual/test_plan.html#thread_group> which says: The ramp-up period tells JMeter how long to take to "ramp-up" to the full number of threads chosen. If 10 threads are used, and the ramp-up period is 100 seconds, then JMeter will take 100 seconds to get all 10 threads up and running. Each thread will start 10 (100/10) seconds after the previous thread was begun. If there are 30 threads and a ramp-up period of 120 seconds, then each successive thread will be delayed by 4 seconds. Ramp-up needs to be long enough to avoid too large a work-load at the start of a test, and short enough that the last threads start running before the first ones finish (unless one wants that to happen). Start with Ramp-up = number of threads and adjust up or down as needed. You can also take a look at JMeter Ramp-Up Period: The Ultimate Guide <https://www.blazemeter.com/blog/jmeter-ramp-up-period> 2. Normally 3 main metrics <https://jmeter.apache.org/usermanual/glossary.html> are: concurrency (number of active threads), throughput (number of requests per second) and response time. You can use whatever data you can obtain to enrich it like resources usage on the system under test side, APM tools metrics, etc. 3. I'm not familiar with these terms 4. The main idea is that your test needs to be realistic, to wit each JMeter thread must represent either a real user using a real browser <https://medium.com/@malith.jayasinghe/performance-testing-with-a-think-time-64b6b737e3f9> or an upstream system thread, real users need some time to "think" between operations and this needs to be reflected in your test. Upstream systems also normally doing polling periodically so I would rather suggest using think times. --------------------------------------------------------------------- To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org For additional commands, e-mail: user-h...@jmeter.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org For additional commands, e-mail: user-h...@jmeter.apache.org
