Poker Training Platform – Observability with OpenTelemetry

This project enhances the observability of a poker training microservices platform (checkraise.fr) using OpenTelemetry, Grafana, and its associated stack (Loki, Grafana, Tempo, Prometheus).
It provides a complete view of logs, traces, and metrics across all services to analyze system behavior, detect bottlenecks, and improve reliability.

Install

1. Clone the repository:

git clone

2. Install dependencies for all microservices + frontend:

cd apps/auth-service && npm install
cd ../problems-service && npm install
cd ../lives-service && npm install
cd ../../frontend && npm install

Usage

1. Run the dockers for observability stack and the microservices:

cd infra
docker-compose up -d --build

2. Run the Next.js frontend locally:

cd ../frontend
npm run dev

3. Access the application: Open your browser and navigate to http://localhost:3001 to access the poker training platform.

4. Access Grafana dashboard: Open your browser and navigate to http://localhost:3000 to access the Grafana dashboard.

   • Default credentials: admin / admin (you will be prompted to change the password on first login)

---

Architecture

Services:

• auth-service → Handles authentication and JWT management
• problems-service → Manages poker problem creation and retrieval
• lives-service → Implements a gamified “lives” system

Frontend: Next.js (runs locally)
Backend: Node.js + Express + Prisma + PostgreSQL
Deployment: Docker Compose for all observability components

apps/                           # Microservices
├── auth-service
├── problems-service
└── lives-service
frontend/                       # Next.js app (runs locally)
infra/                          # Observability and docker setup
├── Grafana/  
        └── tp1_open...json     # Pre-configured Grafana dashboard
├── OpenTelemetry Collector/
├── Loki/
├── Tempo/
├── Prometheus/
└── docker-compose.yml

OpenTelemetry Integration

Each microservice integrates OpenTelemetry through a tracing.ts file.

• Auto-instrumentations: HTTP, Express, Prisma
• Exporters:
  • Logs → OTLPLogExporter → Collector → Loki
  • Traces → OTLPTraceExporter → Collector → Tempo
  • Metrics → OTLPMetricExporter → Collector → Prometheus
• Correlation: Every log includes trace_id and span_id for cross-navigation in Grafana.

Example Log (JSON structured):

{
  "timestamp": "2025-10-25T14:22:33Z",
  "level": "info",
  "service_name": "auth-service",
  "message": "User authenticated successfully",
  "trace_id": "af23d8c1f5...",
  "span_id": "be09c3..."
}