Unanswered question

Many JavaScript actions - performance issues?

Has using of many JavaScript actions a significant negative impact on the performance of test runs? Some have concerns that to load and use about 20 js actions for about 350 virtual users for some hours will distort this tests.

I'm sure that this js scripts get loaded at the start of a scenario, and not for each user in each iteration. Do you have a statement on this, or would you advice against usage of many small js script actions? Is there a way to find out when those jscripts take too much resources?

** Background **: We have about 15 UserPath's in a NL project, and all pathes have to run together in a scenario. We will use about 450 virtual users and have 4 load generators on 4 machines.

We use about 15 js actions in each UP, about 5 of them are in a SharedContainer, which every UP uses after each http-post. They are for errorhandling (if error messages are in the html response), logging and setting variables different for each UserPath, so that SharedContainer can get used independently from a single UP.

We also have 3 JS libraries, where we have most of functions we use frequently. Functions in the libraries are about 10 to 20 lines, using context.variableManager

The js action code snippets are short, about 5 to 15 code lines, and 90% of them is setting and getting variables.

Roland D.
Roland D.

Roland D.

Level
0
40 / 100
points

Answers

Neil W.
Neil W.

Neil W.

Level
4
5000 / 5000
points
Team

My understanding is that these don't effect performance or timing metrics. The impact will be the number of VUs that a LG can run. The more complex the userpath, the bigger its footprint. As you only have 450 users spread across 4 LGs, it's only 112 users per box so their should be plenty of CPU/RAM available