Reading the Resource Usage graphs: averages, limits and faults in plain words

cPanel & WHM·· 2 views· 6 min read

The short answer: each Resource Usage graph has three things: the average (the usage curve), the limit (the line marking your account’s ceiling) and the faults (the times usage hit the ceiling and the account was held back). Normal is a curve well below the line with no faults. Close to the limit is a curve that touches the line at its peaks, with rare faults whose cause you know. A problem is faults on ordinary days, or a curve stuck to the line for hours. How to reach the graphs is in checking resource usage in cPanel.

The lines and the columns, one by one

What you see What it means How to read it
Average The average usage in each interval of the period The main curve. It shows how the account usually is, but hides short peaks.
Max, in the table view The highest point within the interval If the maximum touches the limit and the average does not, there were peaks: check the faults.
Limit Your account’s ceiling for that resource The line you compare everything against. A graph with no line, or marked unlimited, has no ceiling for your account.
Faults How many times usage reached the ceiling and was held back Zero faults: the account was never held back in that period, however high the curve looks. This is the number that matters most.
MySQL (on some servers only) The share of CPU and I/O used by the database If the MySQL share is the largest, the cause is in the queries: slow queries.

What a fault does to the site, resource by resource

Resource What the visitor notices during a fault
CPU The site slows down but still answers. There is no error, so nobody suspects it.
Memory The request fails: a blank page, a 500 or 503 error, an import that stops halfway.
Entry processes (EP) Extra requests are turned away: a 508 error page (Resource Limit Is Reached) or a 503, and seconds later all is well again.
Number of processes (NPROC) Scripts, cron jobs and Terminal commands fail to start.
I/O and IOPS The site is slow until the heavy disk operation finishes.

Each error is explained in website errors explained.

Reading a graph, step by step

1 Pick the right period. For an incident, the shortest period that contains it (the last hour, the last 4 hours, the last 24). For a trend, 7 or 30 days. Over a long period, each point is the average of a bigger interval and a peak of a few minutes gets flattened.
2 Read the faults before the curve. If there are no faults, the account was not held back, however high the curve looks. If there are, note the resource and the time.
3 Compare the curve with the limit line. Far from the line is headroom. Touching it at the peaks is “close to the limit”. Stuck to the line for hours is the account working at the ceiling. CPU is shown as a percentage: compare it with the line, not with 100% of a whole machine.
4 Look at the other graphs at the same time. Two faults in the same minute usually share a cause: two limits reached at once.
5 Recognise the pattern with the table below: a peak at the same time every day is not read like a curve that climbs slowly.
6 Look for the cause at that time: the Snapshots, where available, the Errors page and the access logs (watching your site usage in real time over SSH).
7 After you change something, read the same period again a week later. If the faults have gone, it worked.

The most common patterns

What you see on the graph What it usually is
A low curve, no faults Normal. Nothing to do.
A peak at the same time every day A scheduled task: a backup, an import, a cron job (cron jobs).
A curve stuck to the line for hours, at any time of day Something running non-stop (a stuck script, a loop) or robots crawling the site (AI bots).
EP faults with low CPU Slow requests waiting on something else: the database or an outside service. They hold slots without using CPU.
Memory faults at the same time as 500 errors A plugin, an import or a report asking for more memory than the account has.
CPU and EP rise together with visits and drop at night Real usage. If the faults come at the busiest hours, that is growth.

Optimise or upgrade?

What you see Optimise or upgrade?
Faults in a daily peak at the same time Optimise. Move the task to another time or split it into parts.
A curve stuck to the line even in the small hours Optimise. Something is stuck, or it is robots: too many processes running.
EP faults with low CPU Optimise. The requests are slow: hitting the entry processes limit.
Faults only at the busiest hours, and more each month Upgrade. The site has grown.
You have optimised, the average dropped, and faults continue at peak hours Upgrade.

Bigger plans

On a bigger plan the limit line sits higher and the same curve gains headroom. We do not publish CPU, memory, process or I/O figures for each plan; the bigger plans have more of all of them, from Pro to Plus and to Ultra, and the Business line for shops and sites with traffic peaks. After the upgrade, compare the same period before and after: fewer faults is the proof. How to do it: how to upgrade your plan; which step to pick: which hosting plan comes next.

Do not trust the average on its own. A low average with faults means short peaks that have already shown visitors an error. The opposite misleads too: a single peak on a 30-day graph, with no faults, is not a problem. Decide by the faults, over the right period.
Put the EP graph next to the CPU one. If both rise together, the requests are working: caching and fewer robots help. If EP hits the limit while CPU stays low, the requests are waiting: look for slow queries or an outside service that is slow to answer.

Not sure what the graph is telling you? Open a support ticket with a screenshot of the period and we will tell you what is hitting the ceiling.

Open a support ticket

SEE ALSO

Checking resource usage in cPanel: CPU, memory, processes and I/O for your account

Your plan limits at a glance: where to see each one in cPanel and in your account

CPU, RAM and disk I/O in plain words: what each one limits on your site

Entry processes: why a site with few visitors can still hit the limit

Website errors explained: 500, 403, 404, 508 and what to do

Two limits reached at once: which resource to fix first and when an upgrade solves both

RECOMMENDED PRODUCT

Web hosting with cPanel

Domain and SSL included, daily backups and the panel you already know. from $6.60/mo (3-year plan, with coupon)

See plans
Was this answer helpful? Your answer helps us improve this article.

Related Articles

cPanel & WHM 13,845 How to access cPanel Control Panel? cPanel is the control panel where you manage everything in your hosting account - files, e-mail accounts, databases, bac...
cPanel & WHM 2,042 How to create an e-mail account in cPanel? Creating a professional e-mail address on your own domain - such as info@yourdomain.com - takes less than a minute in cP...
cPanel & WHM 2,468 How to change your cPanel password There are two ways to change your cPanel password, and the difference is simple: one needs the current one, the other do...
cPanel & WHM 1,681 What can you do in cPanel? cPanel is the control panel included with your hosting plan. Everything about your site - files, e-mail, databases, doma...
cPanel & WHM 1,610 Your cPanel access details To log in to cPanel you need three things: the address, your username and your password. This page explains what each on...
cPanel & WHM 1 Your plan limits at a glance: where to see each one in cPanel and in your account The short answer: your plan’s limits come in two kinds. Some have a published figure: disk space, the number of sites an...