5. Technologies

This chapter summarizes the technologies used to implement, test, and deploy SmartGym Monitor.

5.1 Backend Stack

5.1.1 Java 21

All backend services are built with Java 21. The Gradle toolchain is configured to use this version consistently across modules.

5.1.2 Spring Boot 3.4.3

Spring Boot is the foundation of the backend microservices. It provides the HTTP layer, dependency injection, configuration management, actuator endpoints, and application bootstrap.

5.1.3 Spring Cloud 2024.0.1

Spring Cloud is used for service discovery and API gateway integration. It enables the microservices to register on Eureka and be routed through the gateway.

5.1.4 Spring Web, WebFlux, and Actuator

• Spring Web is used by the synchronous REST services.
• Spring WebFlux supports reactive and non-blocking integrations where needed.
• Actuator exposes health and metrics endpoints used by the Docker health checks and observability tooling.

5.1.5 Spring Data MongoDB

Each microservice persists its own data in MongoDB through Spring Data MongoDB. This supports the database-per-service approach adopted in the architecture.

5.1.6 Eureka

Eureka is the service registry used by service-discovery and consumed by the other backend services.

5.1.7 JWT Security

JWT tokens are issued by auth-service and validated by the gateway. The implementation uses the jjwt libraries and a shared secret configured through environment variables.

5.1.8 Vert.x

Vert.x is used alongside Spring Boot in specific services. The embedded-service relies on the Vert.x MQTT client to connect to the Mosquitto broker, subscribe to device topics, and publish commands. Vert.x was chosen for this role because of its lightweight, non-blocking approach to MQTT communication.

5.2 Data Stores and Messaging

5.2.1 MongoDB

MongoDB is the main persistence technology. Separate databases are used for auth, area, machine, tracking, analytics, and embedded data.

5.2.2 MQTT / Mosquitto

The simulator publishes events to Mosquitto, a lightweight MQTT broker, and the embedded service consumes them. MQTT is the main protocol for device-to-backend communication.

5.3 Frontend Stack

5.3.1 Python 3.12

The dashboard frontend is implemented in Python 3.12.

5.3.2 Flask 3.0.3

Flask is used to build the lightweight web interface for login and dashboard views.

5.3.3 Gunicorn

Gunicorn is the production-ready WSGI server used to run the Flask application.

5.4 Simulator Stack

5.4.1 Go 1.24

The device simulator is implemented in Go. It is small, fast, and well suited to generating background traffic in a long-running process.

5.4.2 Paho MQTT Client

The simulator relies on github.com/eclipse/paho.mqtt.golang to publish MQTT messages.

5.5 Testing and Quality Assurance

5.5.1 JUnit and Cucumber

The backend repository includes JUnit tests and a Cucumber-based e2e suite.

5.5.2 Pytest

The Flask frontend is tested with pytest.

5.6 Tooling and DevOps

5.6.1 Gradle

Gradle orchestrates the Java backend build. It manages dependencies across the multi-module project, using a Version Catalog to centralize versions and ensure consistency.

5.6.2 Poetry

Poetry manages Python dependencies and the Flask application environment.

5.6.3 Go Modules

Go Modules handle the dependencies for the Go simulator, ensuring reproducible builds and version management.

5.6.4 Docker and Docker Compose

Docker is used to containerize each component, while Docker Compose orchestrates the full stack locally.

5.6.5 VitePress and Mermaid

The documentation site is built with VitePress, and Mermaid is used for diagrams.

5.6.6 GitHub Actions

GitHub Actions implements the CI, documentation deployment, release, and commit validation workflows.

5.6.7 Husky, commitlint, and Renovate

• Husky enforces local Git hooks.
• Commitlint validates Conventional Commits.
• Renovate automates dependency update pull requests.

5.6.8 Semantic release

Semantic Release handles automated releases from the main branch.