Skip to content

topology.md

Latest commit

 

History

History
123 lines (98 loc) · 3.68 KB

topology.md

File metadata and controls

123 lines (98 loc) · 3.68 KB
Raw

Deployment Topology

Docker Compose (Local Dev)

The local development stack runs five services on a shared bridge network (careguard). Persistent data is kept in named Docker volumes.

graph TB
  subgraph external["External Dependencies"]
    H["Horizon RPC<br/>(soroban-testnet.stellar.org)"]
    OZ["OZ Facilitator<br/>(channels.openzeppelin.com)"]
    LLM["LLM Provider<br/>(api.groq.com)"]
  end

  subgraph compose["Docker Compose — careguard network"]
    subgraph services["Services"]
      S["server<br/>port 3004<br/>volume: ./data"]
      D["dashboard<br/>port 3000<br/>Next.js"]
      R["redis:7-alpine<br/>port 6379<br/>volume: redis-data"]
      P["prometheus<br/>port 9090<br/>volume: prometheus-data"]
      G["grafana<br/>port 3030<br/>volume: grafana-data"]
    end

    S -->|health check| D
    S -->|rate limit / cache| R
    S -->|metrics| P
    P -->|data source| G
    S -->|HTTP| H
    S -->|HTTP| OZ
    S -->|HTTP| LLM
  end

  style external fill:#f5f5f5,stroke:#999,stroke-dasharray:5 5
  style compose fill:#e8f4f8,stroke:#333
Loading

Port Map

Service Internal Port Host Port
server 3004 3004
dashboard 3000 3000
redis 6379 6379
prometheus 9090 9090
grafana 3000 3030

Persistent Volumes

Volume Mount Purpose
./data /app/data Server data (spending history, audit logs)
redis-data /data Redis append-only persistence
prometheus-data /prometheus Time-series metrics (35d retention)
grafana-data /var/lib/grafana Dashboard configuration

Render (Production)

The production deployment runs a single web service on Render's Node.js runtime. Monitoring and caching services are not self-hosted in production; the server relies on managed Redis (Render Redis) and external monitoring providers.

graph TB
  subgraph external["External Dependencies"]
    H["Horizon RPC<br/>(soroban-testnet.stellar.org)"]
    OZ["OZ Facilitator<br/>(x402/testnet)"]
    LLM["LLM Provider<br/>(api.groq.com / llama-3.3-70b)"]
    RENDER_REDIS["Render Redis<br/>(managed)"]
  end

  subgraph render["Render — careguard-api"]
    S["server (Node.js 22)<br/>port 3004<br/>start: node dist/server.js"]
  end

  subgraph clients["Clients"]
    W["Dashboard (Next.js)<br/>hosted separately"]
  end

  W -->|HTTP :3004| S
  S -->|cache / rate-limit| RENDER_REDIS
  S -->|HTTP| H
  S -->|HTTP| OZ
  S -->|HTTP| LLM

  style external fill:#f5f5f5,stroke:#999,stroke-dasharray:5 5
  style render fill:#fef3e2,stroke:#333
Loading

Render Service Properties

Property Value
Type web
Runtime Node.js 22
Plan Free
Build npm ci && npm run build
Start node dist/server.js
Health /health

Resource Comparison

Resource Docker Compose Render
Server Container on careguard network Single web service
Dashboard Container on careguard network Hosted separately
Redis redis:7-alpine container Managed Render Redis
Prometheus prom/prometheus container External monitoring
Grafana grafana/grafana container External monitoring
Persistent storage Named volumes Render disk (if configured)

Service Dependencies

  • server → redis (caching, rate limiting)
  • server → prometheus (metrics export)
  • dashboard → server (REST API)
  • server → Horizon RPC (Stellar on-chain queries)
  • server → OZ Facilitator (x402 payment channels)
  • server → LLM Provider (AI-powered interactions)