Skip to content

QUICKSTART.md

Latest commit

 

History

History
100 lines (68 loc) · 2.51 KB

QUICKSTART.md

File metadata and controls

100 lines (68 loc) · 2.51 KB
Raw

Sentinel AI — Quick Start

Start the Hub (2 minutes)

cd sentinel_ai
kill $(lsof -ti :5001) 2>/dev/null; pkill -f "python.*main.py" 2>/dev/null
source venv/bin/activate
brew services start ollama
python main.py

Open http://localhost:5001

Connect the Raspberry Pi

SSH into the Pi and run:

python3 sentinel_client.py --hub http://<HUB_IP>:5001 --device raspberry-pi-ECE510

Or start it permanently via systemd (auto-restarts on crash/reboot):

systemctl --user start sentinel-client

The Pi streams CPU, memory, disk, network, and ESP32 IoT sensor data (soil moisture, pump voltage, relay state) to the hub every 5 seconds.

View IoT Sensor Data

  1. Go to dashboard → Distributed Devices & IoT Nodes tab
  2. Click raspberry-pi-ECE510 in the sidebar
  3. Scroll up to the 🌱 IoT Sensors — ESP32 card

Readings update every 2 seconds:

  • Soil Moisture % — color-coded: green (wet), yellow (moderate), red (dry)
  • Pump Voltage — supply voltage from the external 5V PSU
  • Relay / Pump — ON or OFF
  • Soil ADC Raw — raw 12-bit ADC reading
  • Moisture progress bar

Control the Pump

In the IoT Sensors card at the bottom:

Button Action
▶ ON Forces pump relay ON for 5 minutes
■ OFF Forces pump relay OFF for 5 minutes

After the 5-minute override, the ESP32 returns to autonomous mode: soil < 40% → pump ON, soil ≥ 40% → pump OFF.

Run a Stress Test (Full Pipeline Demo)

# Trigger CPU spike on hub machine
curl -X POST http://localhost:5001/api/simulate/start/cpu_spike

# Trigger on Pi
curl -X POST http://localhost:5001/api/devices/raspberry-pi-ECE510/queue_command \
  -H "Content-Type: application/json" -d '{"action":"stress_cpu"}'

# Watch the pipeline fire:
# Anomaly → AI Diagnosis → Recovery actions
# Check: http://localhost:5001 → Activity Log

Within 30–60 seconds you'll see:

  1. Anomaly detected — CPU/memory spike flagged
  2. AI Diagnosis — Groq LLM identifies root cause
  3. Recovery — algorithmic fix applied (renice, throttle, kill, purge)

Power Sag Test (IoT Power Monitoring)

curl -X POST http://localhost:5001/api/simulate/start/power_sag

Simulates a -0.75V voltage drop on the hub's simulated power metrics. The anomaly detector fires a critical power anomaly within one detection cycle.

Stop Everything

curl -X POST http://localhost:5001/api/simulate/stop
pkill -f "python.*main.py"