1. The best way is to simulate these users actions as close as possible. The main point is that real users don't hammer the system under test non-stop, they need some time to "think" between operations so make sure to carefully and properly implement these think times in your tests <https://medium.com/@malith.jayasinghe/performance-testing-with-a-think-time-64b6b737e3f9>.
2. When it comes to execution, most probably you will need to at least amend JMeter's configuration, the default setup is good for tests development and debugging but most probably won't suit for high loads so make sure to follow JMeter Best Practices <https://jmeter.apache.org/usermanual/best-practices.html> 3. Also remember that JMeter needs some headroom to operate as if it will not be able to send the requests fast enough you will get false negative results with lower throughput that the system under test can actually handle so get used to monitor the baseline resources usage (like CPU, RAM, Disk and Network IO) - any of them shouldn't be higher than i.e. 90% of maximum available capacity. The resources can be monitored using i.e. JMeter PerfMon Plugin <https://www.blazemeter.com/blog/how-monitor-your-server-health-performance-during-jmeter-load-test> 4. If you can kick off 20000 users using single JMeter machine after step 3 - you should be good to go, if not - see how many users were online when resources consumption started exceeding acceptable thresholds using i.e. Active Threads Over Time plugin, divide 20000 by that number, allocate that many machines of the same hardware specifications and set up a distributed JMeter test <https://jmeter.apache.org/usermanual/remote-test.html> On 6/9/2022 1:52 PM, Deepak Chaudhari wrote:
Hi, I want to execute a load test with 20000 virtual users using JMeter. What is the best way to achieve this? Thanks in advance. Deepak.
