I need to run a load test with 10k users and would need to request about 10 api. What is the recommended system configuration i need to have? If I am using aws to run the script, which instance I would need to use? Is there any tweak I need to do with regards to jvm or anything?
Thanks in advance
Answer depends on what the virtual users are doing, and how fast the users are working. Figuring out the optimal value for your workload won’t be that hard, and shouldn’t take more than a day of experimentation:
Start with one, minimal config. If you run out of memory, you’ll either run out right away, or before the load level reaches its peak.
Otherwise, monitor the box, and watch its CPU usage. If it maxes out, you know you need more CPU cores.
If you get errors about being unable to establish a connection, then you know that you need a second injector box.
If you find yourself in that boat, you might want to invest in Gatling FrontLine, which will take away the headache around managing a cluster of Gatling injectors. I’m using it. It comes highly recommended. 