Add Prometheus metrics endpoint for HTTP observability

[kskarimi]

This PR adds a Prometheus scrape endpoint to improve black-box monitoring visibility for Kutt HTTP traffic.

What changed

• Added a new /metrics endpoint (configurable via METRICS_PATH) that returns Prometheus text format.
• Added request metrics middleware to capture:
  • Request duration histogram (http_server_requests_seconds)
  • Request counters (total/success/error)
  • In-flight requests
  • Throughput gauges (RPS/RPM)
  • Per-status-code request counts and rate metrics
• Added new env vars:
  • METRICS_ENABLED (default: true)
  • METRICS_PATH (default: /metrics)
• Updated .example.env and README with metrics configuration and Kubernetes scrape annotation example.

Why

This provides operational visibility for:

• Response time
• Throughput
• Error/success behavior by HTTP status code
• Prometheus-based scraping in Kubernetes environments

Kubernetes scrape annotations example

prometheus.io/scrape: "true"
prometheus.io/path: "/metrics"
prometheus.io/port: "3000"

Notes

• Metrics are in-memory per app instance and reset on process restart.
• Suitable for per-pod scraping in Kubernetes.