Skip to content

Cross-cutting Concepts

Dummy Text for section 8:

This section describes overarching concepts and solution patterns that are used throughout the IsoPrüfi system to ensure consistency, reusability, and architectural integrity.

Domain Concepts

  • Insulation Performance (ΔT): The key metric is the temperature difference between indoor and outdoor sensors. This value is computed in real time and used for analysis and alerts.
  • Sensor Units: Each Arduino is considered a logical sensor unit, tagged with a physical location and ID.
  • Data Sources: Each data record is enriched with a source type (sensor, weatherAPI), a timestamp, and a confidence level (based on fallback or delay).

Safety and Fault Tolerance Concepts

  • Redundant Logging: Each Arduino writes to local SD cards to preserve data if the network fails. Data can later be synchronized.
  • Retained MQTT Messages: MQTT topics are configured with retain=true so the latest sensor values are cached on the broker.
  • Stateless Services: Backend services are designed to be stateless and independently deployable, which simplifies recovery after failure.

Architecture and Design Patterns

  • Message-Driven Architecture: Core communication is asynchronous via MQTT. Services subscribe to well-defined topics using a publish/subscribe model.
  • Microservice Pattern: Backend services (WeatherWorker, MQTTReceiver, API) are independently deployable and loosely coupled.
  • API Gateway Pattern: The Website Cluster communicates with backend services through the REST API cluster, shielding internal complexity.

Development Concepts

  • Containerization: All services run in Docker containers and are grouped logically (e.g., "MQTT Cluster", "Website Cluster").
  • Continuous Integration: GitHub workflows enforce formatting, build checks and tests before merges.
  • Infrastructure as Code: Deployment manifests are version-controlled (e.g. docker-compose.yml, Kubernetes manifests).

Operational Concepts

  • Observability via Loki: Logs from all services are collected via the Loki log system and visualized in Grafana.
  • Health Monitoring: Each service exposes a /health endpoint and reports metrics via Prometheus-compatible format (optional).
  • Scaling: Stateless services behind a load balancer can be scaled horizontally (e.g., Website Cluster, MQTTReceiver Cluster).

These cross-cutting concepts ensure consistency across all layers of the system and support the project’s core quality goals: persistence, availability, data integrity, and maintainability.