That’s a really hard question to answer without knowing more about your infrastructure setup as well as your test scripts. I’d suggest you login to one of your runners over ssh and run your test script while monitoring the CPU and RAM saturation of the runner. You also mention vcpus, which make the calculation even harder, as the saturation of the rest of your cluster will also impact your result.
If your test script only consists of pure ES5.1+ compliant javascript, and you dont do any expensive dynamic data generation as part of your scripts, I’d say that 100 VUs probably would fit. But again, without testing it’s quite hard to say for sure.
For me to find out Infrastructure SetUp and get access of Runner and monitor performance of it will be time taking activity. I need to contact several people.
So what I did. I set up ubuntu EC2 instances.
1 - t2-micro-1vcpu-1gbram-100vu
Setup K6. Executed test suite with 100 VU. It takes around 45 minutes. Avg CPU consumption is 70% till test suite runs.
2 - t2-medium-2vcpu-4gbram-100vu
Setup K6. Executed test suite with 100 VU. It takes around 45 minutes. Avg CPU consumption is 40% till test suite runs.
We have plan to execute test cases with 200 and more users in future. Currently, w can not run due to data dependency.
Therefore, from above results, can I conclude that t2-micro-1vcpu-1gbram-100vu might create problems if VU load increases. CPU might be heavily utilize. Responses might get slow.