select(
                "generateDiffs",
                "",
                display = ParameterDisplay.PROMPT,
                description = "Skips differential flame graph generation to avoid OOM in large adhoc tests.",
                options =
                    listOf(
                        "Generate diffs" to "",
                        "Skip diffs generation" to "--no-generate-diffs",

        flameGraphGenerationBuildSpec.run {
            PerformanceTest(
                model,
                stage,
                flameGraphGenerationBuildSpec,
                description =
                    "Flame graphs with $profiler for ${performanceScenario.scenario.scenario} | ${performanceScenario.testProject} " +
                        "on ${os.asName()} (bucket $bucketIndex)",
                performanceSubProject = "performance",

crearíamos ese objeto para cada request, y estaríamos leyendo el archivo `.env` para cada request. ⚠️

Pero como estamos usando el decorador `@lru_cache` encima, el objeto `Settings` se creará solo una vez, la primera vez que se llame. ✔️

{* ../../docs_src/settings/app03_an_py310/main.py hl[1,11] *}