Uptime monitoring: how to know when your site is down

An uptime monitor is a service that requests your pages from several places every one to five minutes and alerts you by email, text or chat app when they fail to load. For a company website, set one up on the homepage and the contact page, with a keyword check so a broken page counts as down, plus alerts for your SSL certificate and domain expiry. It takes about fifteen minutes, and a free plan covers most company sites.

The point is simple: you should find out your site is down before a customer tells you. Without a monitor, the usual way businesses learn about an outage is a phone call, days later. Monitoring is part of every website care and maintenance plan we run, with alerts going to us and to a named person at the client, but you can set up the basics yourself with the steps below.

What does an uptime monitor actually check?

Different checks catch different failures. Most services offer all of these:

If a status code in an alert is unfamiliar, the HTTP status code reference explains each one in plain English.

The certificate check matters more than it used to. Let's Encrypt, which issues a large share of the web's certificates, stopped sending expiry reminder emails in June 2025, and certificate lifetimes are getting shorter. A renewal that fails now gives you less warning. SSL certificates explained covers why.

Which pages should you monitor?

Start small. A monitor on every page creates noise, and noise gets ignored.

• The homepage, because it is the page most people land on and the one that fails when the whole site fails.
• The contact page, because it is the page that earns money for most company sites.
• Any checkout, booking or login page, if you have one.
• The form or API endpoint that receives enquiries, if it lives on a different server from the pages.

For each one, pick a keyword that only appears when the page renders properly: a heading or a phrase from the footer, not the company name, which also appears on most error pages.

How do you set it up?

1. Pick a service. UptimeRobot, Better Stack and similar services offer free plans that check every five minutes, as of October 2026. Uptime Kuma is a free open-source option you host yourself, which only makes sense if it runs somewhere other than the site it watches. Read the terms of any free plan; some restrict commercial use.
2. Add an HTTP monitor with a keyword for each page on your list. Set the interval to five minutes, or one minute for a store or app.
3. Require confirmation from more than one location before an alert fires, if the service offers it. This removes most false alarms caused by a single network hiccup.
4. Turn on SSL and domain expiry alerts, set to warn at least 14 days ahead.
5. Add two alert contacts. One person at your company and one at your agency or developer is a good pair. Use email plus one channel someone will see at a weekend, such as a text message or a chat app.
6. Break it on purpose. Change the keyword to something that is not on the page, wait for the alert, and confirm both people got it. Then change it back.

If you run scheduled jobs, such as nightly backups or a feed import, add a heartbeat monitor that expects a ping after each run. The cron expression generator helps you write and check the schedule the job and the monitor should share.

If you already pay someone for care, ask who receives their alerts and how fast they act on one. An alert that lands in an unread shared inbox is the same as no monitor.

What should happen when an alert fires?

Decide this before the first alert, not during it. A one-page plan is enough:

1. Confirm it. Load the site on mobile data, not the office network, and check the monitor's details for the error and the location.
2. Check the obvious causes. The host's status page, the domain's expiry date, the certificate's expiry date, and any change made in the last 24 hours.
3. Call the person who can fix it. Have their number written down, along with the host's support link and the registrar login location.
4. Write down what happened. Time down, time back, cause and fix. Three lines is enough, and after a year it shows whether the same thing keeps failing.

What does an uptime percentage actually mean?

Hosts quote uptime as a percentage. The numbers sound alike and are not:

For a company website, 99.9% is a fine target. Paying extra for 99.99% hosting rarely makes sense for a brochure site; most of its real outages come from expired domains, broken plugins and bad deployments, which no hosting guarantee covers. Spend the money on monitoring and someone who responds instead.

What does monitoring not tell you?

A monitor sees that a page loads. It cannot tell you that the contact form's emails are landing in spam, that a payment gateway is rejecting cards, or that the site loads in nine seconds. It also misses hacks that hide spam pages from normal visitors.

So add a monthly human check. Someone submits every form and confirms the email arrives in the right inbox. If form emails go missing, the cause is usually email authentication, covered in SPF, DKIM and DMARC. If Google starts warning about your site, your website got hacked: a recovery plan is the next read.

Monitoring is the cheapest part of the maintenance routine in the website maintenance guide, and the one to set up first. Do the homepage and contact page today, and add the expiry checks before you close the tab.