What problem does it solve? Software systems often suffer from hidden performance bottlenecks, slow user flows, and resource inefficiencies that are hard to identify without a systematic, measurement-first methodology. ## Core Features & Use Cases - Bottleneck Identification: Systematic profiling, tracing, and log analysis to detect CPU spikes, memory leaks, high latency, and lock contention. - Full-Stack Optimization Guidance: Covers Core Web Vitals (LCP, INP, CLS) for frontend, plus database indexing, N+1 query avoidance, caching, and connection pooling for backend. - Observability & Verification Cycle: Uses OpenTelemetry, Prometheus, Grafana, and distributed tracing with a Measure-Analyze-Optimize-Verify loop including regression testing. - Use Case: When an API endpoint responds slowly, follow the workflow to baseline current latency, profile the code, form an optimization hypothesis, apply one atomic change, and verify improvement with before/after metrics. ## Quick Start Ask the agent to analyze a slow endpoint or page, establish baseline performance metrics, and propose a verified optimization plan.
