I try to measure our responsetimes and started simple with one endpoint.
The request is targetting an amazon Application load blancer.
Thereby I noticed that the first request in a scenario has a delay. The delay is allways there, regardless what I try.
I have read the documentation that such short living virtual Users are not ideal, but that shareConnections should mitigate this. And that warmUp (either in code as below or in the conf file) should also hide this delay from the results.
I test with gatling 3.5.1 and have tested this on Mac Big Sur, Windows 10 and Ubuntu subsystem on Windows 10.
import io.gatling.core.Predef._
import io.gatling.http.Predef._
import scala.concurrent.duration._
class ReproducerRunner extends Simulation {
var httpProtocol = http.baseUrl(“https://integration.messebackend.aws.corussoft.de”)
httpProtocol.warmUp(“https://integration.messebackend.aws.corussoft.de/ok”) // Changes nothing?
httpProtocol.shareConnections // Changes nothing?
object Warmup {
val warmup = exec(
http(“Warmup request to static response”)
.get("/ok")
)
val warmup2 = exec(
http(“Warmup request to static response 2”)
.get("/ok")
)
}
setUp(
scenario(“Still warmup time”)
.exec(Warmup.warmup)
.exec(Warmup.warmup2)
.inject(
constantUsersPerSec(10) during (10 seconds)
)
).protocols(httpProtocol)
}