Resolving High Load Average Issues

Understanding Load Average

uptime
# 10:30:00 up 45 days, load average: 4.50, 3.20, 2.10

The three numbers represent 1-minute, 5-minute, and 15-minute averages. A load of 1.0 means one CPU core is fully utilized. On a 4-core server, a load of 4.0 means all cores are busy.

Diagnosis

# Is it CPU, I/O, or both?
top
# Look at: %us (user CPU), %sy (system CPU), %wa (I/O wait)

# Detailed CPU breakdown
mpstat -P ALL 1 5

# I/O statistics
iostat -x 1 5

High CPU Load

# Find CPU-hungry processes
ps aux --sort=-%cpu | head -10

# Check for runaway processes
top -b -n 1 | awk 'NR>7 && $9>50 {print $1, $9, $11}'

High I/O Wait

# Find processes doing heavy I/O
iotop -o

# Check disk queue depth
iostat -x 1 | awk '$1=="sda" || $1=="nvme0n1p1\" {print "await:", $10, "queue:", $9}'

Quick Fixes

• Runaway process — kill it: kill -TERM PID
• Too many PHP workers — reduce pm.max_children
• Database queries — check slow query log, add indexes
• Log writing storm — verbose logging can cause I/O pressure