Set up monitoring with Prometheus

Set up Prometheus monitoring for a [Language/Framework] application. Expose a /metrics endpoint and instrument the following signals: HTTP request duration as a histogram, error rate as a counter, and active in-flight connections as a gauge. Write a complete Prometheus scrape configuration. Then create three Grafana dashboard panels for the most important operational signals and add alerting rules for p99 latency exceeding [Xms] and error rate exceeding [X%].