uptime (for example load average: 0.81, 0.50, 0.47) and compare it with your core count from nproc. A load equal to the number of cores means the CPUs are fully busy; consistently higher means work is queuing.
Before you start
- Connect with SSH. None of the commands here need
sudo. - Know your core count. On a VPS this is the number of vCPUs in your plan;
nprocshows what the server sees.
How to check the load average
Any of these shows the same three numbers:top, htop and w show the same values in their header line.
What do the three numbers mean?
The kernel counts tasks in stateR (running or ready to run) and state D (uninterruptible sleep, usually waiting for disk). It turns that count into three rolling averages in which recent samples weigh more than older ones.
Read them as a trend:
Because they are averages, they react slowly. In our test, running one CPU-bound process on an idle 2-core server for 20 seconds moved the 1-minute value from
0.23 to 0.81, not to 1.00.
What is a high load average?
The numbers are not divided by the number of CPUs. Theuptime manual gives the example: a load of 1 means a single-CPU system is busy all the time, but a 4-CPU system was idle 75% of the time.
There is no single “bad” number. A load of 4 on a 4-core build server can be fine; on a game server it may cause lag. Watch for sustained values above your core count, and a 15-minute value that keeps climbing.
Is load average the same as CPU usage?
No. CPU usage is the share of time the CPU was busy. Load average also counts tasks waiting for disk (stateD). So you can see a load of 10 with the CPU mostly idle if the disk is the bottleneck.
Check which it is:
vmstat don’t show the last column (gu, guest time).
Ignore the first line of
vmstat output; it is an average since boot. The next lines are live.
What to do when the load is high
1
Confirm it is sustained
Look at the 5 and 15-minute values. A short 1-minute spike during backups or updates is normal.
2
Find the cause
High
r and us: a busy program. See find what is using CPU.
High b and wa: disk. See check disk I/O.
High si/so: the server is swapping. See check memory usage.3
Fix or resize
Stop or optimise the busy program, limit its CPU, or move to a plan with more cores or RAM. A full walk-through is in fix high load.
Frequently asked questions
What is a good load average for a 2-core VPS?
Below2 most of the time. Short peaks above that are fine; sustained values of 3 to 4 or more mean work is queuing.
Why is my load high but CPU usage low?
Tasks are waiting for disk or network storage (stateD). Run vmstat 1 5 and check b and wa, then iostat -xz 1 to see which disk.
Does load average include threads?
Yes. Linux counts each thread as a scheduling entity, so a busy multi-threaded program can add several points to the load on its own.Why does the load average change after I stop a process?
The numbers are rolling averages. The 1-minute value drops within a minute or two; the 15-minute value takes much longer.Related guides
VPS and dedicated server guides
Every server guide in one place.
Fix high load
Find and fix the bottleneck.
Monitor resources with htop
Live CPU, memory and processes.
Find what is using CPU
Track down a busy process.
Check disk I/O
iostat and iotop.
Checked against the uptime(1), proc_loadavg(5) and vmstat(8) manuals on 4 November 2026; commands tested on Ubuntu 24.04 LTS. Commands can differ on other distributions or releases.
Need Extra Help?
If you encounter any issues, our support team is ready to assist:- Live Chat: Quick assistance via our website.
- Support Ticket: Open a Ticket
- Discord: Join our Community
- Email: support@freakhosting.com
Last Updated: November 2026