Cost‑efficient injection for low‑traffic Gatling scenarios

Cost‑efficient injection for low‑traffic Gatling scenarios

The objective is to lower cloud instance‑hour expenses for Gatling tests that simulate a low‑traffic workload, such as one user generating requests at one per second.

Constraints include using Gatling 3.10+ DSL helpers, accounting for JVM warm‑up and garbage‑collection pauses that may affect resource measurements, and ensuring response‑time variance stays within acceptable bounds.

Which injection profile—constantUsersPerSec or rampUsers—provides the lowest cost while maintaining acceptable response‑time variance in low‑traffic scenarios? How can cloud provider metrics be used to verify cost savings without being biased by JVM warm‑up or GC effects?