Part 3: The Insight — Catching the Bottleneck Red-Handed

Part 3 of a series on AI-assisted performance engineering demonstrates how to diagnose a database connection pool exhaustion that caused a 34.18% error rate and 89.9-second latency spike during a 1,000-user stress test. A Python integration layer fuses JMeter JSON reports with HikariPool metrics and feeds them into GPT-4o with a 'Senior Performance Architect' prompt. The AI correlates the data and delivers a definitive root cause verdict: resource starvation from a saturated 100-connection pool, recommending increasing max-pool-size to 250 and implementing a Circuit Breaker pattern. The next part will integrate this flow into a CI/CD pipeline for autonomous performance engineering.