Performance Testing of a Mendix App with JMeter

0
Hi,   As we're currently trying to implement a performance test solution for our Mendix based application using JMeter we're facing a challenge when it comes to two of the payload/request body attributes from the /xas endpoint.  These request body attributes are "changes" and "objects". We're facing a correlation issue due to a very high dynamicity of data that is put under these two attributes. For e.g. we've noticed that at some point under "changes" attribute for one of the /xas POST request there can be even more than 1400 lines. The data from these 1400 lines is dynamically put by Mendix as I've discussed with our developers and they mentioned that is not under their control. Trying to do a correlation in this case is a nightmare as the data used under "changes" or "objects" attributes is fetched from various different previous /xas requests' response. There's no actual rule we could identify and of course we did not investigate the whole 1400 lines as it's not productive at all.    I was able to do the correlation for some /xas requests which only had one or two objects under the "changes" attributes and the request works. But for such case where we have this large amount of data is a nightmare and we couldn't find an approach.   You can see an example in the attached screenshot. Pay attention to "changes" and "objects" attributes.   Is there a way we can build the data that goes into those two attributes using Jmeter? Is the data from "changes" and "objects" really a must?   I've tried to send empty objects but I get a 506 response code when doing the /xas request. I get the same 506 response code if I use the default values that I've recorded via Jmeter. This tells me that new data is being used with each run for these attributes and  I need to to the correlation otherwise the request won't work.  
asked
1 answers
0

Due to the complexity of performance testing in Mendix by simulating the xas requests especially via JMeter, it could be better to just not do it. 

 

I think you're bettor off by trying to use free tools for monitoring performance like the nanoprofiler. 

 

https://marketplace.mendix.com/link/component/118285

 

You could also check slow running queries by looking into the logs of you application for the following messages 

 

WARNING - ConnectionBus_Queries: Query executed inWARNING - ConnectionBus_Queries: Query executed in ... 

 

This will let you know which SQL queries took how much to execute. 

 

Hope this helps!

answered