Skip to content

Load Testing

A load test is what every step before this one has been preparing for. You take a test case that replays cleanly for one user and you scale it to hundreds or thousands of concurrent users, hold that load long enough to see how the system behaves, and look at the metrics. The point isn't to make the system fail. The point is to find out where it would, before real users do it for you.

In v7.0 you can generate load locally, bound only by your CPU and bandwidth, with no built-in virtual-user cap on the Free tier. Or you can distribute the generators across AWS EC2 in any combination of regions. The cloud path uses AWS SDK v2 and provisions engines faster than v6.x did.

In This Section

  • Concepts: Virtual users, think time, ramp-up patterns, load profiles, and the difference between a stress test and a soak test. Skim this if you've run load tests before. Read it if you haven't.
  • Configuring a Load Test: Defining the load profile. User counts, ramp-up time, duration, think-time scaling, and which datasets feed which users.
  • Cloud Load Testing: Setting up AWS credentials, choosing regions, sizing engines, and launching distributed tests across multiple regions.
  • Running a Load Test: Starting the test, what you'll see while it runs, and the controls (pause, halt, abort) you have during execution.
  • Monitoring During Tests: What to watch in real time. Response times, error rates, virtual user counts, server health. AI-driven alerts surface here when something starts to degrade.

If you've never run a load test before, work through Concepts first. The rest of the section assumes you know what virtual users and think time mean. Once that's in place, Configuring a Load Test is the next step.