We have to do load testing for the qlik application using mashups, application uses the web sockets for communication. Please let us know if Jmeter can be used to test the application ?
Using LoadRunner I was able to record and replay the web socket calls but challenge is LoadRunner is not waiting to receive the response for the web sockets calls, instead loadrunnre keeps on sending the socket requests.
Please let me know if anyone tested qlik mashp up successfully and captured the response times when filters are applied in the dashboard using either LoadRunner, Jmeter or Neoload.
In the thread Scalability Tools vs. JMeter for Performance testing for Mashups? its mentioned that web sockets with Jmeter is an obstacle.
Sebastian Fredenberg Please let me if its feasible to test with Jmeter as 80%of actions (more than 5000 socket calls) performed uses web socket communication
If you want to test with jmeter (or any other load generating tool) you need to make sure that you have a websocket implementation added as it is not supported by default in at least jmeter. It is a tricky task to handle, that is why we recommend to use scalability tools to test Sense.
The problem with Scalability tools is that it does not support mashups (as there is no way to tell how the mashup is implemented). However, you can use the scalability tools to load test the application holding the data that the mashup is using, just make sure that you access a sheet containing all objects that would appear in the mashup.
Thank you for your responses.
I have two questions,
1.- Wich load generating tool are scalibility tools now based on?
2.- Is it posible to test mashups with load runner? If yes, can you give us deep information on how to add the correct websocket implementation to test qlik sense with load runner?
Thank you very much
1. The Sense scalability tool use the .Net SDK provided by Qlik. The Qlikview Scalability tool use jmeter.
2. It should be possible, as Loadrunner supports websockets. I have not used loadrunner with websockets so I cannot help out how to do that. As with any load generating tool I assume that it would require quite some work to get it running.
Thank you for your quick response.
Sorry for my ignorance about this and for my insistence...
You mentioned above that "The problem with Scalability tools is that it does not support mashups (as there is no way to tell how the mashup is implemented)". Is this because a limitation of Scalability tools? or Is it a limitation because of using mashups, independently of the testing tool used?
I ask this because I need to know any way to test mashups. Then if load runner can do that, I'll need some practices to integrate with, or at least some information for understanding the way qlik sense uses websockets. Anybody worked on this way?
I really thank your help
I can put it this way: The Sense scalability tools are specifically tailored to communicate with Qlik Sense, sending and recieving the same communication that a real user would do using a browser. This in itself is a great strength (and one of the reasons we promote the scalability tools over other tools - communication will be performed correctly). Another strength is the simplicity of creating a scenario to simulate, so it will be fast and communication will be performed correctly.
A mashup communicates with Sense objects but it can be created in many different ways that is unknown to the tool, and also exists on a web page which of course looks different to others. Thus, there are no predefined "mashup action" that behaves in a specific way.
So, you can simulate mashup usage by creating an app with the same objects as in the mashup and interacting with that using scalability tools. Or you need to create the logic and communication from scratch using another load generating tool.
Thank you Sebastian.