Enhanced Prometheus Metrics

[mftee]

Extend the existing MetricsRegistry with histogram and gauge metrics to provide richer observability over the service's performance and state.

Requirements:

• Add the following metrics to MetricsRegistry:
  • request_duration_seconds (Histogram with buckets: 0.005, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1.0, 2.5) — HTTP request latency
  • stellar_query_duration_seconds (Histogram) — Stellar Horizon query latency
  • cache_size_entries (Gauge) — approximate number of entries in Redis cache
  • active_connections (Gauge) — number of active HTTP connections
  • circuit_breaker_state (Gauge with label state) — 0=closed, 1=half-open, 2=open
• Implement MetricsRegistry::observe_request_duration(duration: f64) and observe_stellar_duration(duration: f64)
• Add Axum middleware that measures request duration and calls observe_request_duration after the response is sent
• All metrics must have descriptive help strings

Acceptance Criteria:

• GET /metrics returns valid Prometheus text format including the new histograms and gauges
• Histogram buckets are correctly configured
• Middleware correctly measures end-to-end request time
• Unit tests verify metrics are incremented/observed correctly