feat: add circuit breaker, error handler, Prometheus metrics, and Dockerfile improvements

[Odkinggjnr]

Description

This PR implements four production-readiness improvements for the Heliobond backend:

1. Circuit Breaker for Stellar RPC Calls (Closes #200)

• Added CIRCUIT_BREAKER_THRESHOLD and CIRCUIT_BREAKER_COOLDOWN_MS configurable environment variables (with backward-compatible fallbacks to existing RPC_BREAKER_* vars)
• Replaced console.warn with structured logger.warn for circuit state transitions (CLOSED → OPEN → HALF_OPEN), including metadata for observability
• Updated .env.example with new circuit breaker env var documentation
• Added tests verifying logger integration and configurable thresholds

2. Express Error-Handling Middleware (Closes #199)

• Updated error handler to use structured logger.error for server-side logging instead of console.error
• Changed error code to INTERNAL_ERROR (uppercase) matching the issue's acceptance criteria format: {"error":{"code":"INTERNAL_ERROR","message":"..."}}
• Stack traces are logged server-side but never exposed to clients
• Added test for async route handler error catching (Express 5 compatibility)
• Updated existing tests to validate the new error response format

3. Prometheus Metrics Endpoint (Closes #201)

• Installed prom-client and created src/lib/prometheus.ts with metrics registry
• Added /metrics endpoint returning Prometheus text format
• HTTP metrics: http_request_duration_seconds histogram and http_requests_total counter (labeled by method, route, status_code)
• Stellar RPC metrics: stellar_rpc_call_duration_seconds histogram and stellar_rpc_calls_total counter (labeled by operation, result)
• Cron metrics: cron_job_duration_seconds histogram and cron_job_runs_total counter (labeled by job, result)
• Transaction metrics: stellar_tx_submissions_total counter and circuit_breaker_state gauge
• Added prometheusMiddleware for automatic HTTP request instrumentation
• Comprehensive test suite for the /metrics endpoint

4. Dockerfile for Containerized Deployment (Closes #198)

• Updated to Node.js 20 alpine (per README specification)
• Fixed production port to 3001 (matching config.PORT default)
• Added non-root user (appuser) for security best practice
• Multi-stage build (build + production) for optimized image size
• Updated docker-compose.yml with required env vars (ADMIN_SECRET_KEY, PROJECT_REGISTRY_CONTRACT_ID)
• Expanded .dockerignore to exclude test files, docs, and dev configs
• Added tests for multi-stage build, port 3001, and Node.js 20 base image

Type of change

• New feature
• Bug fix
• Breaking change
• Documentation update

Testing checklist

• Tests pass locally
• Added new tests (if applicable)
• Code follows project conventions

Related issues

Closes #198, Closes #199, Closes #200, Closes #201

Notes

• The bun.lock file may need to be regenerated with bun install to include the new prom-client dependency before CI passes with --frozen-lockfile
• All env var additions are backward-compatible with sensible defaults

[drips-wave]

@Odkinggjnr Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits