Hi All,
I am new to the Gatling and I started implantation of load testing for my project. I want to run the test for 10M or more constant active virtual users in the system. I am searching for a proper way to do this type of testing so could anybody please help here to identify the proper injection profile from which I could do the test.
Thanks,
Regards,
Ratnesh.
Something like:
inject (
rampConcurrentUsers(100) to(10000000) during(10 minutes),
constantConcurrentUsers(10000000) during(1 hour))
but you won’t be able to generate this kind of load without resorting to distributed tests, typically with FrontLine.
Hi Stéphane,
Thanks for your quick response, Before using any type of Injections profile I have a question-related workload system as mention in Gatling documentation. There are two types of workload systems open and close and my workload system is open workload system so here is my question is could I use a close injection profile for an open workload system?.
Thanks & Regards,
Ratnesh Mishra.
No, it doesn’t make any sense. You would generate artificial conditions that would only exist in the lab and you wouldn’t be able to draw any conclusions on how your system will behave when facing real (open workload) traffic.
The only situation where it can temporarily make sense is change management when having a long history with another tool that emphasizes closed workload model and switching to Gatling and wanting to keep your baseline during migration, and switch to open workload model in the next step.
Thanks, Stéphane for the help. I think then I can use your suggested injection profile.
Hi Stephane,
I am running the below simulation but Active Users in the system at any time reach about 1K only but not 100K as I am expecting.
setUp( constantConcurrentScenario.inject
rampConcurrentUsers(100) to (100000) during (20 minutes),
constantConcurrentUsers( 100000 ) during ( 1 hour )))
Whereby using some other injection profile, I am running the simulation for 1000K and Active Users in the system at any time reach about 1000K.
Thanks & Regards,
Ratnesh Mishra.
Active users != concurrent users.
Please read the documentation: https://gatling.io/docs/current/general/reports/#active-users-over-time
Concurrent users metric is available in FrontLine.
I don’t think we can send 100K users per second so is there any other alternative way where we can execute100K Active Users along with the simulation.
concurrent users ~= new users/sec * average journey duration
You have to properly craft users’ journeys.
