FIND YOUR LEVEL ON THE COMPLEXITY CURVE.

1. The curve has thirteen steps, from “one script, one engineer” to “more time on the tool than on the tests”. Each step lights up when your answers match the condition shown in its card. Your level is the furthest step that applies.
2. Tests grows with simulations, services, teams, protocols and test types, and adds weight when requests aren’t shared, test data is hard-coded, or targets aren’t parameterized.
3. Infrastructure grows with network constraints (internal-only targets, IP allowlisting, CI runners that can’t reach the target, code that must stay in-house), peak users, self-built injectors, regions, environments and production testing. A single Gatling load generator handles up to 60,000 concurrent virtual users, so peaks above that need several.
4. Results grows when runs live in separate reports or home-built dashboards, when assertions don’t fail the build, and when stress or soak tests have no history to compare against.
5. Overall complexity weights tests at 40%, infrastructure at 35% and results at 25%.

What this is and isn’t

The curve is illustrative, taken from Gatling’s build vs buy journey. Scores are rules of thumb to start a conversation, not benchmarks. The layout and pipeline follow the project structure in Writing realistic tests and the CI guides for GitHub Actions and GitLab CI. Check plugin and image versions against your project before you commit the files.