🧠 Backend Tech Stack & Architecture - Monlai Project

This backend powers real-time FCM push notifications, async processing, and microservices for driver/customer engagement on Monlai. Built with scalability, observability, and modularity in mind.

---

🚀 Tech Stack

🏗 Core Frameworks & Language

• Go (Golang) – Backend language for all microservices
• Gin – Lightweight HTTP framework
• Uber Fx – Dependency injection framework for modular service composition

---

🧱 Microservices

Each of the following services runs independently in Docker containers:

• notify – Handles Firebase push notification logic
• driver – Driver management microservice
• customer – Customer-related microservice

---

💡 Architecture Patterns & Advanced Concepts

🔄 Asynchronous Processing

• AWS SQS: Queue-based decoupling for driver and user notification triggers
• Concurrent Workers: Fixed goroutine worker pool listens on channel notifyChan

🔁 Redis Patterns

• Token Storage: FCM tokens for each user/driver are stored using role-based Redis keys
• (Optional) Redis Strategy Pattern:
  • Used via strategy.GetTokenStrategy(role) abstraction
  • Enables plug-and-play support for customer/driver-specific token handling
• Redis Pub/Sub / Locking (if implemented in extensions):
  • Considered for distributed coordination and event notification (pending code trace)

---

📦 Message Queuing

• Amazon SQS
  • DRIVER_QUEUE_URL for messages targeting drivers
  • USER_QUEUE_URL for messages targeting users
  • Ensures loose coupling and retry durability
  • Messages deleted after successful enqueue or delivery

---

📲 Push Notifications

• Firebase Cloud Messaging (FCM) – Delivered via the Go Admin SDK
• Per-user token targeting – Dynamically fetched from Redis
• Dynamic message composition – Structured payloads built per role/message context

---

📈 Observability

• Prometheus Client (Go) – Metrics exposed on /notify/metrics
• Grafana Dashboard – Connected via Prometheus scrape config to visualize:
  • Memory usage
  • Go runtime stats
  • Queue handler latency
• API Gateway Proxying – AWS Gateway routes public traffic to private /metrics for Prometheus

---

🔐 Secrets & Credential Management

• Firebase Service Account JSON – Injected securely into container via volume
• .env file – Centralized configuration for:
  • AWS credentials
  • Queue URLs
  • Redis URL
  • Firebase JSON path

---

🐳 Deployment

• Docker: Multi-container deployment (driver, customer, notify)
• Docker Compose: Local orchestration
• GitHub Actions:
  • CI/CD pipeline: builds, pushes to DockerHub, and SSH deploys to EC2
  • Writes service account and .env dynamically

---

🌐 Infrastructure

• Amazon EC2 – Hosts all Docker containers
• API Gateway (AWS) – Publicly exposes FCM registration and metrics endpoints
• Redis – In-memory caching & token store (hosted or containerized)

---