Using Selenium for load testing is not recommended. Whereas JMeter is well suited for high-load performance tests, especially through BlazeMeter, Selenium is more suited for UI functional tests.
This is primarily because each "user" in a Selenium test runs through its own individual browser, which means if you're running 10 users on 1 engine, you are launching 10 web browsers simultaneously. Launching many multiple browser instances is a major CPU consumption per engine.
For running Selenium tests on BlazeMeter, we recommend using Taurus and starting with a maximum of 5 users per engine, then adjust accordingly as resources allow (which you can monitor via the Engine Health tab of your test report). Additional guidance on how to properly calibrate a test for best performance can be found in Calibrating a Taurus Test.
There is a way around the recommended five-user maximum, however: You can execute a multi-test in which each single test runs no more than five users each. You can add the single test multiple times to the multi-test. For example, if you wanted to test a fifteen-user load, you could create a multi-test that includes three instances of the single five-user test.
