But please note there are some http reqs that have no extensions so if we only give the above extensions only then the pages with no extensions will not be recorded. So we have to be very careful with the include list.
View results tree and View results table listeners require lots of memory, so use them only debugging puposes but for large scale samples don’t use them.
In aggregate graph we can save the results to an external file. Try saving file as .csv So that it requires less space and uncheck the xml option while saving the file.
For running long running tests to save memory for bsh sampler click Reset bsh.Interpreter before each call .
In thread group ramp-up period should not be 0.Because otherwise all threads will be bombarded to server at the same time. Also never put ramp-up period too high as well.
Also try to put number of Threads(users) per machine(per pc) less than 300 to avoid jmeter related bottlenecks.
Also in real time environment add a Gausian random timer to http script recorder while recording and for that put the deviation to 300 and constant delay offset equal to ${T}.This will record the actual delay time it takes to reach various pages of the websites.
We can do distribute testing i.e not run more than 300 threads per machine and do rest of the testing on other machines.
If we see our machine is not able to handle lots of threads we can go to the jmeter bin folder and look for jmeter.bat file open in notepad and search for heap and try to increase the heap size.
Another good practice in jmeter is to Add cookie Manger to http requests if we are going to run a lof them.