Linux System Metrics You Should Track

Why Monitor?

Proactive monitoring catches problems before they cause outages. Here are the essential metrics every server administrator should track.

CPU Metrics

• Load average — uptime shows 1, 5, and 15-minute averages. Values above your CPU core count indicate saturation.
• CPU utilization — use mpstat -P ALL 1 to see per-core usage
• IOWait — high iowait means processes are waiting for disk I/O

Memory Metrics

• Available memory — free -h shows total, used, and available. Watch "available" not "free" (Linux uses free RAM for caching)
• Swap usage — any swap activity on an SSD server indicates memory pressure
• OOM kills — check dmesg | grep -i "out of memory"

Disk Metrics

• Space usage — df -h shows filesystem usage. Alert at 80%
• Inode usage — df -i. Running out of inodes prevents creating new files even with space available
• I/O throughput — iostat -x 1 shows read/write rates and queue depth

Network Metrics

• Bandwidth — nload or iftop for real-time traffic
• Connection count — ss -s shows connection statistics
• Packet errors — ip -s link shows error counters

Quick Health Check Script

#!/bin/bash
echo "=== Load ===" && uptime
echo "=== Memory ===" && free -h
echo "=== Disk ===" && df -h /
echo "=== Top Processes ===" && ps aux --sort=-%mem | head -6