DRIFTGATE

DRIFTGATE is an API contract governance and runtime reliability platform built around a single pipeline:

flowchart TD
  A["Angular 20 UI"] --> B["Node.js/Fastify webhook gateway"]
  B --> C["FastAPI runtime guard"]
  C --> D["PostgreSQL contract registry + outbox"]
  D --> E["MongoDB raw payload / document history"]
  E --> F["LangGraph review workflow"]
  F --> G["Azure Service Bus-compatible delivery adapter"]
  G --> H["Prometheus / Grafana metrics + k6 benchmarks"]

Loading

Tech Stack

• Angular 20 UI for the operator console
• Node.js and Fastify for HMAC-verified webhook ingress and idempotency
• FastAPI runtime guard for live payload tracking and drift classification
• PostgreSQL for the contract registry, version history, subscriptions, and transactional outbox
• MongoDB for raw payload captures, diff evidence, validation errors, and document history
• LangGraph for grounded contract review workflows
• Azure Service Bus-compatible delivery adapter for drift-event delivery
• Prometheus and Grafana for metrics and dashboards
• k6 for load and benchmark runs

Services

• frontend/: Angular 20 control room
• gateway/: Node.js/Fastify webhook gateway
• app/: FastAPI runtime guard, review workflow, event publishing, and document-store integration
• backend/: scheduled monitor and changelog service

Quickstart

docker compose up -d --build
curl -X POST http://localhost:8301/api/monitor/run-once \
  -H "X-DRIFTGATE-ADMIN-SECRET: dev-secret"

• Frontend: http://localhost:5173/
• Monitor API: http://localhost:8301
• Runtime guard API: http://localhost:8302
• Gateway: http://localhost:8303

Runtime Behavior

• Webhook ingress is verified with HMAC and idempotency checks in the gateway.
• The runtime guard records payload snapshots, schema diffs, validation failures, and review evidence.
• PostgreSQL stores registry metadata and outbox state.
• MongoDB stores raw payload documents and history.
• LangGraph powers the contract review workflow.
• The delivery adapter can target an Azure Service Bus-compatible sender.
• Metrics are exposed for Prometheus and visualized in Grafana.
• k6 benchmarks are captured as JSON artifacts under docs/benchmarks/.

Local Development

docker compose up -d --build

Useful environment defaults:

• EVENT_BACKEND=noop for local development
• EVENT_BACKEND=azure_service_bus when an Azure Service Bus sender is injected
• DOCUMENT_STORE_BACKEND=memory for tests and minimal local mode
• DOCUMENT_STORE_BACKEND=mongo with DOCUMENT_STORE_URI=mongodb://mongo:27017 for local MongoDB

Notes

• The repository uses the DRIFTGATE project name throughout the codebase.
• Benchmarks, simulations, and generated artifacts live in JSON or code form instead of markdown docs.
• The architecture is designed to be local-first and cloud-compatible without requiring cloud deployment.