Prometheus /metrics endpoint for API monitoring

[Cbiux]

Summary

Add a /metrics Prometheus endpoint for production monitoring of API health, request rates, and scoring throughput.

---

Background

/health confirms process status. Operators need metrics for Grafana/Datadog: request latency, error rates, events ingested per platform, Horizon call failures.

---

Proposed implementation

Endpoint: GET /metrics (Prometheus text format)

Metrics (initial set):

Metric	Type	Description
zcore_http_requests_total	counter	By method, route, status
zcore_events_reported_total	counter	By platform, eventType
zcore_horizon_errors_total	counter	Horizon fetch failures
zcore_auth_verify_failures_total	counter	SEP-53 verify failures
zcore_db_query_duration_seconds	histogram	Prisma query latency

Middleware: Express middleware wrapping app.use in app.ts

Optional auth: METRICS_SECRET bearer token for production

---

Acceptance criteria

• /metrics returns valid Prometheus exposition format
• Middleware tracks all /api/* routes
• Document scrape config in Docs/vercel-deploy.md
• No high-cardinality labels (no wallet addresses in labels)
• Unit test for metric increment

---

Related

• Add GET /api/health endpoint #3 health endpoint (extended in recent commits)
• Provision production MySQL and configure DATABASE_URL on Vercel #15 production ops

---

Out of scope

• Grafana dashboard JSON
• Vercel-native metrics (use external scraper)

[Errordog2]

/attempt #44

[Cbiux]

Resuelto en 9842a63: GET /metrics Prometheus, contadores HTTP/eventos/Horizon/auth, histograma Prisma.