Skip to content

Introduction to GreenCop

Educational Project Disclaimer

This is an educational project built to demonstrate IoT monitoring architecture, machine learning integration, and cloud-native development practices. Not intended for commercial use.

What is GreenCop?

GreenCop is full-stack observability for physical infrastructure. Monitor server rooms, data centers, and critical hardware with real-time telemetry and predictive AI alerts.

Stop reacting to failures. Start preventing them.

Why GreenCop Exists

A single server room outage costs businesses an average of $5,600 per minute. Traditional monitoring solutions only alert you when it's too late—after temperatures have spiked, after hardware has throttled, after your services have degraded.

GreenCop predicts problems 20 seconds before they happen using machine learning anomaly detection. You get advance warning to take action before conditions become critical.

Key Differentiators

Predictive vs Reactive

Most monitoring tools alert you when thresholds are crossed. By then, damage may already be done.

GreenCop's ML-powered anomaly detection analyzes patterns in your sensor data and warns you about unusual behavior before it escalates. Temperature trending up faster than normal? You'll know before it hits your hard limit.

Tip

Enable prediction feedback after your first week. Mark which alerts were actionable vs. false positives. The model retrains nightly and gets smarter with your specific infrastructure patterns.

Zero-Configuration Deployment

No IP addresses to configure. No manual sensor registration. Power on an ESP32 sensor and it auto-discovers the gateway via mDNS (greencop-gateway.local), registers itself with a unique hardware ID, and starts streaming data.

Setup time: ~2 minutes per sensor.

Dual-Layer Intelligence

Combine threshold-based alerts (immediate hardware risk) with ML anomaly detection (early warning system) for comprehensive coverage.

  • Hard Limits: Temperature > 30°C → instant alert
  • ML Anomalies: Unusual pattern detected → advance warning
  • Predictions: Forecasted threshold breach in 20 seconds → preventive action window

Alert Strategy

Set hard limits 5°C above your comfort zone for true emergencies. Use ML alerts for everything else to reduce noise. Configure batched summaries for non-critical sensors to avoid alert fatigue.

Who Should Use GreenCop?

Data Center Operators

Monitor temperature and humidity across racks, prevent costly downtime, and maintain compliance logs for audits.

Value: Reduce MTTR by 60% with unified monitoring vs. siloed tools.

IT Infrastructure Teams

Prevent hardware failures from environmental issues, reduce downtime, and sleep better knowing predictive alerts are watching 24/7.

Value: One prevented outage pays for the system for an entire year.

DevOps Engineers

Integrate monitoring data into your existing workflow. Trigger automated HVAC adjustments, route alerts to different Teams channels based on sensor location, export compliance reports programmatically.

Value: Full REST API access. Build custom automations in hours, not weeks.

IoT Students & Researchers

Learn cloud-native IoT architecture with real hardware, understand event-driven systems, practice with modern web technologies (React, FastAPI, Terraform).

Value: Portfolio-worthy project demonstrating end-to-end system design.

How It Works

graph TB
    subgraph "Edge Layer"
        S1[IoT Sensors<br/>ESP32 Nodes]
    end

    subgraph "Ingestion Layer"
        GW[Gateway Service<br/>mDNS Auto-Discovery]
        PS1[Pub/Sub Topic: data]
    end

    subgraph "Processing Layer"
        CF1[Cloud Function<br/>Data Ingestion]
        CF2[Cloud Function<br/>Alert Detection]
        ML[ML Service<br/>Anomaly Prediction]
    end

    subgraph "Storage Layer"
        BQ[BigQuery<br/>Time-Series Data]
        DB[PostgreSQL<br/>Metadata]
    end

    subgraph "Application Layer"
        API[FastAPI Backend<br/>Cloud Run]
        FE[React Frontend<br/>Railway]
    end

    S1 -->|Auto-Discover<br/>HTTP POST| GW
    GW -->|Publish Events| PS1
    PS1 -->|Trigger| CF1
    PS1 -->|Trigger| CF2
    CF1 -->|Write| BQ
    CF2 -->|Analyze| ML
    ML -->|Store Predictions| DB
    API <-->|Query| DB
    API <-->|Analytics| BQ
    FE <-->|REST API<br/>JWT Auth| API

Data Flow in 5 Steps:

  1. Sensors measure temperature/humidity every 30 seconds and auto-register via mDNS
  2. Gateway receives data and publishes to Google Cloud Pub/Sub queue
  3. Cloud Functions process events and store in BigQuery
  4. ML Service analyzes patterns and generates predictions
  5. Dashboard displays real-time metrics, alerts, and anomaly predictions

System Components

ESP32 Sensor Nodes (Edge)

  • Low-cost WiFi microcontrollers ($5-10 each)
  • MicroPython firmware for easy programming
  • Auto-registration via mDNS (zero config!)
  • LED status indicators (green: active, red: error)

Hardware Choice

Start with 2-3 ESP32 nodes to validate the system. Scale to hundreds once you've proven ROI. Sensors pay for themselves after preventing one hardware failure.

Go Gateway Service (Local Bridge)

  • High-performance local coordinator written in Go
  • Broadcasts mDNS for sensor auto-discovery
  • Routes data to Google Cloud Pub/Sub
  • Monitors sensor health via heartbeats
  • Zero-config deployment: Sensors find the gateway automatically

Cloud Data Pipeline (Processing)

  • Pub/Sub: Event streaming for real-time data flow
  • Cloud Functions: Serverless processing for data ingestion and alerts
  • BigQuery: Petabyte-scale time-series data warehouse
  • Cloud SQL: PostgreSQL for customer metadata and configurations

FastAPI Backend (Cloud Run)

  • High-performance Python async API
  • RESTful endpoints for all operations
  • JWT authentication with bcrypt password hashing
  • SQLAlchemy ORM with Alembic migrations
  • Full BigQuery integration for analytics queries

React Frontend (Railway)

  • Modern TypeScript SPA with React 19
  • Real-time dashboard updates
  • Interactive Recharts visualizations
  • Multi-language support (English/French)
  • Responsive design for mobile/tablet/desktop

Technology Stack Summary

Frontend: React 19, TypeScript, TailwindCSS, Recharts, Vite Backend: FastAPI (Python), PostgreSQL, Cloud Run IoT Pipeline: ESP32 (MicroPython), Go Gateway, Pub/Sub, Cloud Functions, BigQuery ML: Scikit-learn Isolation Forest, Cloud Run ML service Infrastructure: Terraform IaC, Docker, Google Cloud Platform

Use Cases

Data Center Monitoring

Monitor server room temperatures across multiple locations, receive alerts when cooling systems fail, and maintain compliance logs for SOC 2 / HIPAA audits.

Customer Result: "We caught a failing AC unit 15 minutes before critical threshold. Saved $45K in potential revenue loss." - IT Director, SaaS Company

Laboratory Environments

Ensure research equipment operates within safe environmental ranges, track conditions over time for compliance, and document environmental data for grant reporting.

Storage Facilities

Monitor temperature-sensitive storage (vaccines, electronics, food), prevent product damage from environmental drift, and maintain quality control records.

Educational Projects

Learn cloud-native IoT architecture, practice with real hardware and modern web technologies, build a portfolio-worthy end-to-end system.

Next Steps

Choose Your Path

Ready to explore features?