Not really, you should use the time taken in the full transaction controller ("Open Document" in your case). The open doc request is only one of several requests that happens when a document is opened.
If you are not aIready doing so, I recommend you to use the Analyzer package to look at your test result, it should give a clearer picture.
As for the poll, it is one type of request used in the communication between client and qvs. When a request is fired of (for example a selection in a listbox) all objects that are affected will start to get calculated. Then the client will keep asking the qvs for status until everything is complete, that is the poll. Also, as you noticed, the data will be sent back with the poll and not the selection request - but it will keep on polling until everything is complete.
Thank you Sebastian,
for all your valuable answers and quick responses.
I will be in touch with you if any doubt occurs
Hope you are doing good, with your help my qlik view work is going on well, now I have started working on qlik sense performance testing, on qlik sense mashup dashboards[not qlik sense native app tool] dashboards are installed on to the sever on which I got the access, but when i opened the dashboard report by setting proxy in browser[localhost 8080] and recording the script through Apache Jmeter performance testing tool, login is successful also the layout opened but no report or parameters including graphs are visible, but without proxy its working fine,
please guide me, i have also saw some tutorial using scalability tool[native app] for qliksense but its saying all the time that i need to configure the qlik on the local machine and i have to do the perf test on mashup report not on the qlik sense native app,
There is a Scalability Tool for testing Sense available here on community, however that does simulates going to the qlik sense app, not through a mashup. I want to note that the Scalability Tool for Sense is not jmeterbased as there is limited websocket support so this might be the issue you are running into when recording the script.
My best recomendation when it scomes to testing mashups and sense is to either skip the mashup and test the underlying sense application with the scalability tool, or test going though the mashup but measure response times "manually" using chome dev tools or similar.