Add a monitor

Before you start: you need a URL that is reachable from the public internet. Anything behind a VPN, a firewall or a private network cannot be checked from outside.

Uptime → Add monitor. Checks begin automatically as soon as you save.

The fields

URL

The address to check, for example https://example.com or https://api.example.com/health.

Which URL should I use?

• Your homepage is the obvious choice, and fine.
• A health-check endpoint (often /health or /status) is better if your app has one, because it usually checks the database too — so it fails when the app is truly broken, not just when a page is slow.
• Avoid pages that need a login. They will return an error and look permanently down.

If you leave off https://, we add it.

Monitor name

Optional. Left blank it uses the hostname. Give it a real name if you will have many: Shop — checkout API.

Check interval

1 minute, 5 minutes or 10 minutes.

One minute means you find out fast, and it makes 43,200 requests a month to your site — harmless for most, worth thinking about on a small server or a metered plan. Five minutes is the sensible default for most sites.

Retries before Down

A whole number from 0 to 10, default 3. After a failed check the monitor checks this many more times, one interval apart, and turns Down only if all of them fail too. With 3 at a 1-minute interval that is 4 failed checks in a row, about 4 minutes. 0 marks it Down on the first failure — fastest, but one network hiccup counts as an outage.

Email notifications

On by default. When the monitor turns Down you get one email, then a second with the likely root cause if a server is linked. No more email for that monitor until someone clicks Mark fixed after it is back up. Who receives it is set under Notifications on the Uptime page.

Advanced settings

Hidden behind Advanced settings, because most people never need them.

Request timeout

5, 10, 20 or 30 seconds. Default 10. How long we wait before calling the check a failure.

Raise it if your site is legitimately slow to respond and you are getting false failures. Lower it if you want to hear about slowness as if it were downtime.

Expected status

What counts as success:

Response text check

Optional. The check fails unless the page contains this text.

This is the one advanced setting worth using. A hacked or broken site often still returns a perfectly healthy 200. Checking for a word that only appears when things work — your company name in the footer, "ok" in a health response — catches a site that is technically up and actually wrong.

Keep it short and stable. Do not use text that changes with every deployment.

After saving

The monitor appears in the list and the first check runs within a minute. Read Reading the results for what the numbers mean.

Editing, pausing, deleting

Each monitor's menu offers:

• Check now — run one immediately rather than waiting.
• Pause / Resume — stop checking during planned maintenance so the downtime does not spoil your uptime percentage. Resume checks straight away.
• Edit — change any field.
• Delete — removes the monitor and its history.