Hi Esau;
FWIW: I would setup up jMeter on your development PC to run both PS and the associated PS App(s) so that jMeter can record all the test cases that you need to run. That also includes tweaking the jMeter scripts for concurrent loading as well.
Once the basic jMeter tests are running on the development machine OK, I would then create a clean Azure workspace & transfer (install) PS, PS App(s), DBMS, and jMeter, etc on that *same* Azure instance. That way, there should be little of any network latency. Thus when you are running the jMeter tests against the PS App(s), you would be getting the "best case scenario" if no network latency were encountered. From there, you should be able to estimate what the PS App(s) will actually perform like with the PS App(s) installed on an actual user's PC / VM.
Note that you would *not* be able to run much of a jMeter "load test" with the PS server from the developers machine as the PS License from that environment would only support 5 PB concurrent users. So you would need to change up the PS license to support the number of concurrent users that jMeter is trying to emulate.
HTH
Regards .. Chris
How would running a load test on the same server where the PowerServer is hosted (on Azure) be accurate if it has to run JMeter as well?
That defeats the purpose of an actual load test because its not accurate or as close to production as possible.
Also, wouldnt I need to install PowerBuilder on the server to run the test as explained in the documentation to run the WEB API console?