mich43l/opensips-api repository overview

⁠opensips-apis

opensips-apis is a lightweight service layer that exposes custom HTTP endpoints to support OpenSIPS with special routing-time operations. It centralizes business logic that would otherwise bloat opensips.cfg, keeping SIP routing fast while allowing data-backed decisions and utilities to evolve independently.

---

⁠Overview

• Purpose

  • Offload complex, data-driven, or environment-specific logic from OpenSIPS into a small stateless web service.
  • Enable consistent behavior across environments via configuration (no .env required).

• What it provides (at a glance)

  • Number normalization and formatting utilities.
  • Small policy/utility checks (e.g., simple allow-lists or context guards).
  • Optional database-backed lookups for routing decisions.
  • Health/readiness probes for orchestration.

• How it integrates

  • OpenSIPS calls this service using http_client.
  • Results are returned as compact JSON that can be parsed with OpenSIPS json module.
  • You implement the final routing policy in opensips.cfg.

---

⁠Architecture (High Level)

• Runtime: FastAPI + Uvicorn (Python 3.12)
• State: Stateless application; configuration via environment variables
• Data: Optional PostgreSQL for lookups (indexes recommended on queried fields)
• Packaging: Container-first (Dockerfile included)
• Scalability: Horizontally scalable behind L4/L7 load balancers
• Port: Configurable via PORT (default typically 3000)

---

⁠Configuration (High Level)

• Server: HOST, PORT
• Database (optional): PG* variables for connectivity, plus table/column names for lookups
• Network Controls (optional): Allow-list via comma-separated CIDRs/IPs

---

⁠Deployment (High Level)

• Docker: Build and run the container; map PORT as needed
• Compose/Swarm/K8s: Add health checks, resource limits/requests, and proximity to OpenSIPS/DB
• Resiliency: Configure timeouts/retries/backoff in callers or your gateway

---

⁠OpenSIPS Integration (Conceptual)

1. Load http_client and json modules.
2. From a routing block, POST a small JSON payload (e.g., number/context).
3. Parse JSON response and apply your routing policy (rewrite/route/deny).
4. Log decisions for observability.

---

⁠Observability & Operations

• Health: /health endpoint for liveness/readiness
• Logging: Structured logs for inputs, decisions, timings
• Metrics (roadmap): Prometheus/OpenTelemetry