Expose prometheus metrics via /metrics endpoint

[Allda]

The Python REST application should be able to expose metrics via a public endpoint using Prometheus.

As an application owner, I want to have visibility into several metrics:

• Count of requests per API route
• average, maximum, and minimum duration per route
• HTTP status codes per route

Use standard libraries to expose the metrics.

[fullsend-ai-triage]

Triage Summary

Category: Feature request
Severity: Medium
Status: Ready to implement

Problem

The application currently has no observability into API request metrics — there is no way to monitor request counts, latencies, or error rates per route.

Requested Behavior

• Expose a /metrics endpoint serving Prometheus-format metrics
• Track request count per API route
• Track request duration (avg, min, max) per route
• Track HTTP status codes per route

Recommended Approach

Integrate the prometheus_client library to instrument the Flask app. Add a /metrics endpoint that exposes standard Prometheus text-format metrics including a Counter for request counts (labeled by route and method), a Histogram for request durations (labeled by route), and status code labels. Add the new dependency to requirements.txt and document the endpoint in the README.

Proposed Test Case

1. GET /metrics returns 200 with Prometheus text format (body contains HELP/TYPE directives)
2. After hitting an instrumented route (e.g., GET /items), GET /metrics includes counters and duration observations for that route
3. Status code labels appear in metrics output
4. Edge case: /metrics endpoint itself should not recursively inflate request counters

No duplicate issues or blocking dependencies were found. This is a self-contained feature that can proceed immediately.

---

Labels: This is a new feature request for adding Prometheus metrics, matching the 'enhancement' label.

[Allda]

/fs-code