Simulating large number of users.

[lubms]

I’m trying to understand how the number of users simulation works. Read me mentions that (average of window load + estimated real time user use) / average of window load = real time user number.

I have 20 windows open. Each represents a different page in one report. Each of these pages loads at a different average time. How can I estimate how many average users are using this report if each page will provide a different user account?

Summing all those results seems unrealistic since it will be a very large user count.

My end goal is to simulate around 200 users on one workspace with multiple multi page reports.

[furmangg]

Does the example here help?

For example, if the browser window reports an average of 10 seconds to refresh the report with a 0 seconds think time and you assume a real user will spend approximately 30 seconds studying the report before clicking again, then (10+30)/10 = 4. The load test you just ran simulated 4 "real world users per window".

Ramp up the number of windows until you reach your desired number of simulated real world users and validate performance is adequate. If performance begins to drop then you can record the max number of real world users the report can support before performance degrades.