performance-engineer

End-to-end request tracing identifies hot paths causing p99 latency by applying tracing and diagnostics workflows across services and data paths. It detects tail latency and regressions by comparing established baselines.

Load testing determines capacity headroom and SLOs by exercising systems under representative load. This soak testing reveals saturation scenarios and concurrency bottlenecks to validate fixes with a repeatable test plan.

Reproducible performance fixes require establishing baselines, collecting traces and metrics, and documenting steps. The Skill enforces measurable outcomes by comparing baselines to quantify improvements and validate fixes with a repeatable test plan.

This performance profiling approach focuses on collecting traces and metrics to establish baselines and document reproducible steps. It differs from basic monitoring by applying load-testing workflows to detect p99 latency, regressions, and saturation scenarios.