High CPU or Memory Usage - DigitalFyre Knowledge Base

If your VPS is slow or unresponsive, or showing high resource usage in the Server Statistics, use the steps below to identify the cause.

Identifying the Problem

CPU

# Top processes by CPU usage
top -bn1 -o %CPU | head -20

# Alternative using ps
ps aux --sort=-%cpu | head -15

Look for processes consuming a disproportionate amount of CPU. Common causes include runaway scripts, unoptimised database queries, bot traffic, and compromised processes such as crypto miners.

Memory

# Memory overview
free -h

# Top processes by memory
ps aux --sort=-%mem | head -15

In the free -h output, focus on available memory and swap used. If swap usage is significant, your server is under genuine memory pressure.

Linux uses available memory for disk caching, so high memory usage is not always a problem. The concern is when available memory is very low and swap is being used heavily.

Disk I/O

High I/O wait can appear as high CPU load:

# Check iowait (the "wa" column)
top -bn1 | head -5

A high %iowait value means the CPU is waiting on disk operations rather than doing useful work.

Common Causes and Fixes

Runaway process

# Identify and kill the process
top
kill -9 PID

Check application logs, cron jobs, and scripts for the root cause. Database (MySQL / MariaDB)

# Check active queries
mysql -e "SHOW FULL PROCESSLIST;"

Common fixes: add indexes to frequently queried tables, optimise slow queries, tune the InnoDB buffer pool size. Web server under load

# Check active connections
ss -s

# Check access logs for spikes
tail -f /var/log/nginx/access.log

Consider rate limiting, caching, or blocking abusive IPs if traffic is from bots or an attack. OOM killer events When Linux runs out of memory, the kernel terminates processes to free RAM. Check if this has happened:

dmesg | grep -i oom

If OOM kills are occurring regularly, reduce your application’s memory footprint or upgrade your plan.

When to Upgrade

If your server consistently runs above 80% CPU or memory usage under normal load after optimisation, the workload has outgrown the plan. See Upgrading Your VPS.