Tutorial

How to monitor your website's uptime

Why sites go down, the three ways to know before your visitors do, and the free tools that watch a site around the clock so you do not have to.

Ben Johnson · last checked 2026-09-21 · 4 min read

Uptime is the share of time your website is reachable. Downtime is the rest — the minutes or hours when a visitor gets an error instead of your site. On a free or shared plan there is no service guarantee, so knowing when your site is down, and why, is on you. This page covers how to find out before your visitors do.

The trap everyone falls into is checking the site themselves, seeing it load, and assuming all is well. Your own browser is the worst possible judge. It caches pages, your network caches DNS, and both will happily show you a working site that is broken for everyone else. Real monitoring means testing from outside, on a schedule, whether or not you are watching.

Why sites go down

Downtime is not one problem, it is several, and the fix depends on which you have.

The most common on a busy free plan is a resource limit. Too many visitors arrive at once, the plan’s ceiling is hit, and new visitors get a 508 resource limit error while the site itself is fine. The second is DNS: the name stops pointing at the server, so the site is unreachable even though it is running. That family of failures is covered in the DNS guide. The third is the server or the host itself having a bad day, which is the one thing you cannot fix and can only wait out — and the reason a paid plan with a real uptime commitment can be worth it, as the free versus paid breakdown explains.

The three ways to know

There are three levels of watching a site, from crude to proper.

Check it yourself, from outside. When you suspect the site is down, the first question is whether it is down for everyone or just you. An outside checker answers that in seconds. IsItDown.now loads your site from its own servers and tells you what the rest of the world sees, which immediately separates a real outage from a problem on your own machine.

Set up an automatic monitor. A checker you have to remember to run is no use at three in the morning. A monitoring service tests your site every few minutes, day and night, and messages you the moment it fails. Several run a free tier that covers a handful of sites, which is plenty for one project.

Watch the trend, not just the moment. Good monitors keep a history, so you can see whether your site drops out every day at the same busy hour — a sign you are hitting a resource limit and should add caching — or whether outages are random, which points at the host.

What to check, and how often

Test the real page a visitor loads, not just whether the server answers. A server can be up and still serving a broken page, so a good monitor looks for a word it expects to find on the page, not merely a response.

Every five minutes is a sensible interval for a small site. More often than that rarely tells you anything new and can look like unwanted traffic to your host. Check from more than one location if you can, because a site can be reachable from one part of the world and not another, especially just after a DNS change.

When the monitor says it is down

Do not panic, and do not immediately blame the host. Work through it in order. First, confirm it is really down using an outside checker like the one above, so you are not chasing a problem that only exists on your network. Next, decide whether the name resolves at all: if it does not, it is DNS. If it resolves but the page errors, read the error — a 508 is a resource limit, a 500 is usually your own code or a bad config file, and a plain timeout points at the server. Each of those has its own fix, and knowing which one you are looking at saves you from restarting things that were never broken.

Your uptime plan, in order

For a site you care about:

1. Bookmark an outside checker so you can confirm an outage in seconds.
2. Set up a free monitor that tests the real page every five minutes.
3. Turn on its alerts so it wakes you, not your visitors.
4. Watch the history for a daily pattern that means “add caching.”
5. If outages are random and frequent, that is the host, and it is the signal to consider a plan with a real uptime guarantee.

That is the whole of it. Two of the steps are free tools and the rest is a habit, and together they mean you hear about downtime first.