Understanding CPU, RAM, and disk usage on your server

Before you can fix a performance problem, you need to understand what the numbers mean. This guide explains the key resource metrics on a Linux VPS and which commands to use.

CPU load

Run top or htop to see CPU usage. Key columns:

A load average consistently above your CPU core count indicates the server is overloaded.

RAM and swap

free -m

Output example:

              total  used  free  shared  buff/cache  available
Mem:           1994   890   120      45         984       1059
Swap:           999    12   987

Focus on the available column (not "free")-this is memory actually available for new processes after the OS frees cached memory. If available RAM is consistently near zero, the server needs more RAM or optimization.

Swap use above 100 MB on a 1–2 GB VPS means the server is under memory pressure-processes are paging to disk, causing slowdowns.

Disk space and I/O

# Disk space by mount point
df -h

# Disk I/O stats (install with: apt install sysstat)
iostat -x 1

# Find large directories
du -sh /var/www/* | sort -rh | head -20

In iostat, watch %util-above 80% means the disk is heavily loaded. High disk I/O is often caused by database writes, log files, or a runaway process writing continuously.

Finding the resource hog

Sort processes by CPU or RAM in top: press P to sort by CPU, M to sort by memory. The top process is usually the culprit. Common causes:

• php-fpm-too many PHP processes or a slow WordPress site
• mysqld-slow queries or missing indexes
• python/node-runaway application process
• kswapd-kernel swapping memory to disk (RAM exhaustion)

For MySQL slow queries, enable the slow query log:

SET GLOBAL slow_query_log = 'ON';
SET GLOBAL long_query_time = 1;

Then check /var/log/mysql/mysql-slow.log for queries taking over 1 second.

Related articles