In Kubernetes setup where Gatling execution and server side lies in same cluster we get exact number of requests for each endpoint passed on provided user in setup. However in Pre-prod setup where cluster is different for Gatling execution and Server side code, getting lesser number of requests made as per Gatling report.
For Example for user load of 30 we would get random number of requests for each endpoints of parallel running scenarios like it may be 27 calls for an endpoint A, 25 calls for endpoint B and least being around 10 calls for some endpoints. Also KO is 0 for all them as per report. so what happens to the remaining calls? Also we have provided sufficient resource for Gatling execution like 4 core CPU and 4Gi memory.
Adding code snippet to inject user load setUp(
// parallel execution of scenarios
new myScenario().scenario1.inject(rampUsers(30).during(30 seconds)),
new myScenario().scenario2.inject(nothingFor(1 second), rampUsers(30*3).during(60 seconds)),
new myScenario().scenario3.inject(nothingFor(1 second), rampUsers(30*2).during(60 seconds)),
new myScenario().scenario4.inject(nothingFor(1 second), rampUsers(30).during(30 seconds))
)
Respective Report:
