In this blog post, I would like to share about calculating VUsers for each load injector. How many virtual users will get generated for each of your load injector. Based on the information from various forums, I have written this post.
Before you calculate the number of VUsers for each load injector, you need to determine the single virtual user foot print. Below are the steps to identify the footprint.
1. Design a scenario in the Controller and set the iterations as 25. Run time settings should be configured run until complete.
2. Launch perfmon and set the counters %Processor Time and Private Bytes as shown below.
[auto_thumb width=”150″ height=”150″ link=”” lightbox=”true” align=”center” title=”Calculating VUsers for each Load Injector – Perfmon” alt=”Calculating VUsers for each Load Injector – Perfmon – QAInsights” iframe=”false” frame=”true” crop=”true”]http://qainsights.com/wp-content/uploads/2013/03/Perfmon.jpg[/auto_thumb]
3. To configure the above counters, you need to run the scenario for few minutes.
4. Once you completed the counters set up, run the actual scenario and monitor the data.
5. When the scenario execution completed, open the Analysis and analyse the average CPU and peak memory utilization.
To determine the VUsers for each Load Injector, using below calculation.
Following formula will give the number of VUsers that CPU can sustain:
Number of Vusers per CPU = (70% * Number of Core Processors)/ VUser Average CPU Usage
Following formula will give the number of Vusers that memory can accommodate:
Number of Vusers by Memory = (Total GB RAM of LG – GB RAM allocated for the operating system and other processes) / Vuser Peak Memory Usage
Following are the table with examples:
The number of Vusers that we are looking is the lower of these two values. i.e. “Vusers per CPU” or “Vusers by Memory”.
We’re always posting interesting articles on QAInsights. I request you to subscribe so you don’t miss out anything.