Scalability Tools - Does Change Sheet Method gives the load time?

[naveenbalanagu]

I am planning to do performance testing for a sheet that has 10+ visuals. I am using Change Sheet Method as my first scenario after connection is established to the app.

My questions is  - does the change sheet method accurately report the response time (load time) for the entire sheet (with all the visuals loaded)? Does the response time includes the rendering time as well?

[Daniel_Larsson · Accepted solution]

When doing a change sheet action the tool will do a GetObject, GetProperties, GetLayout and for relevant objects the different GetData requests. This will be done for all objects on the sheet. The response time reported is the time between the first request sent until the last response received. As such it will include network and engine calculation time. The tool does however not execute any javascript and as such does not include client side time such as rending time. This is why we recommend "Complementary Manual Measurements" in the documentation. If you want to see fine grained times for rendering you could e.g. use chrome developer tools to see what time is used where, in firefox there's also a profiler which can be used.

[naveenbalanagu]

Hi Larsson,

Thanks for sharing this information. May I know which documentation has more information on "Complementary Manual Measurements" ? Can you share the document or link to it. Thanks in advance.

[Daniel_Larsson]

I was referring to a small section in the tool documentation (page 27). This section suggests fiddler, but does not go into detail exactly how to do this as many different tools and methods exists depending on required granularity. Using fiddler is one option, others are built in browser tools such as chrome developer tools or profiling tool in firefox which offers very fine grained times. For more automated tests, sikuli or protractor scripts could be written. One more option, if the purpose is to get a feeling of user perceived performance, is to use the tool to ramp up background load, and during this load connect manually to get a feeling of the response times during load. We don't want to be very opinionated of which tools to use for this part.

[naveenbalanagu]

Thank you Larsson. I will try both the tools (Fiddler and Chrome Developer Tools) and see which one works better.