Server Health Dashboard

Know exactly what's happening on your servers. Get alerts before users do.

What You'll Build

A monitoring dashboard that tracks:

• CPU, memory, disk, and network usage
• Process-level resource consumption
• Historical trends and anomaly detection
• Instant Telegram alerts for issues

Requirements

• Plugins: Web Search (for API calls)
• Time: 15 minutes
• Server Access: SSH or monitoring agent installed

Setup

1. Add Servers to Monitor

Monitor these servers:
- prod-web-1 (192.168.1.10)
- prod-api-1 (192.168.1.11)
- prod-db-1 (192.168.1.12)

Check every 1 minute.

2. Configure Alert Thresholds

Alert on Telegram if:
- CPU >80% for 5 minutes
- Memory >90% for 3 minutes
- Disk >85% on any partition
- Network errors detected
- Any server unreachable

3. Set Up Historical Tracking

Store metrics for:
- Last 24 hours (1-minute intervals)
- Last 7 days (5-minute intervals)
- Last 30 days (hourly averages)

Generate weekly performance reports.

4. Process Monitoring

Track top processes by:
- CPU usage
- Memory usage
- Open file descriptors
- Network connections

Alert if any process uses >50% CPU unexpectedly.

Sample Alert

High CPU:

🚨 SERVER ALERT

Server: prod-api-1
Issue: High CPU usage
Current: 87% (5-minute avg)
Normal: 35-50%

🔝 Top Processes:
1. node (PID 1234): 52% CPU
2. postgres (PID 5678): 18% CPU
3. nginx (PID 9012): 9% CPU

💡 Possible causes:
- Traffic spike (check logs)
- Inefficient query
- Background job running

[View detailed metrics →]

Recovery:

✅ ALERT RESOLVED

Server: prod-api-1
CPU: 87% → 42%
Duration: 12 minutes

Root cause analysis:
- Traffic spike from API endpoint /search
- Query optimization deployed
- CPU back to normal

[View incident timeline →]

Sample Weekly Report

📊 Server Health Report — Feb 5-11

🖥️ prod-web-1
Uptime: 99.98%
Avg CPU: 38%
Avg Memory: 62%
Disk: 45% used
Incidents: 0

🖥️ prod-api-1
Uptime: 99.85%
Avg CPU: 52%
Avg Memory: 71%
Disk: 58% used
Incidents: 2 (both auto-resolved)

🖥️ prod-db-1
Uptime: 100%
Avg CPU: 28%
Avg Memory: 78%
Disk: 67% used
Incidents: 0

🎯 Recommendations:
- prod-api-1: Consider scaling to 2 instances (CPU trending up)
- prod-db-1: Memory usage stable but near threshold — monitor
- All servers: Disk usage healthy

Next review: Feb 18, 2024

Pro Tips

1. Predictive Alerts: Use ML to detect anomalies before thresholds are hit
2. Log Correlation: When CPU spikes, auto-pull relevant application logs
3. Auto-Scaling: Trigger cloud provider API to scale up during high load
4. Cost Tracking: Monitor cloud costs alongside performance metrics

---

Monitor your infrastructure → Launch Claws