Create a minimal reproducible performance test to replicate the maximum concurrency issue observed at 200 concurrent users.

Using a lightweight load test, target the primary failing endpoint with a fixed payload and no think time to generate sustained concurrency. The goal is to consistently reproduce the latency spikes and 5xx errors seen under load in the performance environment. Collect server logs, metrics, and connection pool data during the run to identify bottlenecks (e.g., thread pool, DB connections, or network limits). This reproduction will help isolate and diagnose the root cause of the concurrency failure.