System Dashboard

Live system monitoring with web UI, CLI mode, and signal handling.

What is it?

A lightweight Linux system monitor that shows you what's happening on your machine in real-time. Open a browser, check CPU/memory/disk usage, see which processes are hogging resources. Built with Bash and Python, no bloat.

Key points:

• Real-time metrics (CPU, memory, disk, I/O, processes, load, uptime)
• Web dashboard + JSON API
• Signal-based refresh (SIGUSR1)
• Graceful shutdown (SIGINT/SIGTERM)
• Works in terminal too (headless mode)

Options

./run.sh --help

--host HOST       Host to bind (default: 127.0.0.1)
--port PORT       HTTP port (default: 8000)
--interval SEC    Metric refresh interval (default: 2)

Web Interface

Dashboard shows:

• CPU usage + load average
• Memory/swap usage
• Disk I/O rates
• Active processes count
• System uptime
• Per-core CPU breakdown
• Top 5 by CPU and memory
• System alerts (high CPU/memory/disk)

Quick Start

./run.sh

Open http://127.0.0.1:8000 in browser. Done.

Signal Handling

The dashboard handles signals for refresh and graceful shutdown.

run.sh (orchestrator)

Manages both server and monitor processes.

SIGUSR1 — forwards refresh signal to monitor

kill -USR1 $(cat runtime/run.pid)

SIGINT — graceful shutdown (Ctrl+C)

kill -INT $(cat runtime/run.pid)
# Output: Received signal: SIGINT. Stopping…

SIGTERM — terminate gracefully

kill -TERM $(cat runtime/run.pid)
# Output: Received signal: SIGTERM. Stopping…

monitor.sh (metrics collector)

Responds to refresh and shutdown signals independently.

SIGUSR1 — immediate metric refresh (don't wait for interval)

kill -USR1 $(cat runtime/monitor.pid)
# Triggers refresh_all() immediately

SIGINT — exit with status

kill -INT $(cat runtime/monitor.pid)
# Output: Received signal: SIGINT. Exiting…

SIGTERM — exit with status

kill -TERM $(cat runtime/monitor.pid)
# Output: Received signal: SIGTERM. Exiting…

API Endpoints

# Get metrics
curl http://127.0.0.1:8000/api/metrics | jq

# Trigger refresh
curl http://127.0.0.1:8000/api/refresh | jq

# Health check
curl http://127.0.0.1:8000/api/health | jq

Testing Signals

Run the demo

./demo_signals.sh

Shows signal handling in action:

1. Start monitor in background
2. Send SIGUSR1 (refresh)
3. Send SIGINT (shutdown)
4. Show results

Manual test

Terminal 1:

./run.sh --port 8083

Terminal 2:

# Get PIDs
MONITOR_PID=$(cat runtime/monitor.pid)
RUN_PID=$(cat runtime/run.pid)

# Test SIGUSR1 on run.sh (forwards to monitor)
kill -USR1 $RUN_PID

# Test SIGINT on monitor directly
kill -INT $MONITOR_PID

# Or use run.sh orchestrator
kill -INT $RUN_PID

Watch the console output to see which signals are received.

Headless Mode

Run monitor without dashboard (background operation):

./monitor.sh --headless --interval 2 --log-file logs/system_health.log

Writes metrics to files only, no terminal output.

Files

Runtime (auto-generated)

• runtime/monitor.pid — monitor process ID
• runtime/run.pid — orchestrator process ID
• runtime/metrics.env — current system snapshot
• runtime/cpu_cores.csv — per-core CPU usage
• runtime/top_cpu.csv / top_mem.csv — top processes
• runtime/alerts.txt — active alerts
• logs/system_health.log — metric history

Alerts trigger when

• CPU ≥ 85%
• Memory ≥ 80%
• Disk ≥ 85%

Troubleshooting

Port in use?

./run.sh --port 8001

Metrics not updating?

tail -f logs/system_health.log
kill -USR1 $(cat runtime/monitor.pid)

Won't shut down cleanly?

pkill -9 -f monitor.sh
pkill -9 -f server.py

Need to see what's running?

ps aux | grep -E '(run|monitor|server)'