Load Testing
Load Testing is its own section in the navigation rail, separate from functional Runs. A load test runs in one of two modes: Simple — concurrent browser users replay your generated feature tests and time every step — or Advanced — protocol-level k6 load with a configurable profile and pass/fail threshold gates, with the k6 script generated for you. Either way, the run dispatches to the cloud load fleet and charts results live.
What You Can Do
| Task | Use it for |
|---|---|
| Create a load test | Pick journeys, a target URL, and a load profile. |
| Review the generated k6 script | See exactly what will run before you start it. |
| Run under load | Dispatch the test to the cloud load fleet with one click. |
| Watch live results | Latency percentiles, throughput, virtual users, and error rate update while the run executes. |
| Gate on thresholds | Fail the run automatically when a latency or error-rate gate is exceeded. |
| Summarize or export | Ask Assistant for a plain-English summary, or export JSON/CSV results. |
| Automate from CI or AI tools | Create, update, list, dispatch, and inspect load tests over the CLI, REST API, or MCP server. |
How It Relates To Functional Runs
Functional runs check a journey against its assertions. Load tests replay the same journeys — alongside optional protocol-level traffic — many times in parallel to measure behavior under the configured load and check your chosen assertions and thresholds. Journeys need a generated test before they can be used in a load test.