One of the easiest uptime mistakes is assuming HTTP 200 means "everything is fine".
It does not.
Your server can return 200 while showing a blank page, a generic error shell, stale HTML, or the wrong response body. From a simple status-code check, all of those look healthy.
What keyword monitoring changes
Keyword monitoring adds a second condition to the check: not only must the request succeed, the response must also contain text that proves the expected page or API payload is really there.
Examples:
- A dashboard page should contain
Welcome back. - A public status page should contain your service name.
- A health endpoint should contain
"status":"ok".
When this catches real problems
- A deployment returns the app shell but not the expected content.
- A reverse proxy serves the wrong upstream page.
- An auth redirect or maintenance page returns 200 by mistake.
- A third-party dependency fails and your fallback page still responds successfully.
How to choose a good keyword
- Use text that should always be present when the page is healthy.
- Avoid words that can disappear during normal copy changes.
- Prefer machine-stable values for APIs, such as
"status":"ok". - Do not use a keyword that also appears on your error page.
How to set it up in Uptraq
Create or edit a monitor, open its advanced settings, and define the keyword that must appear in the response body. If the request returns 200 but the keyword is missing, the monitor can still mark the check as failed.
Bottom line
Status codes tell you whether a server answered. Keyword checks help tell you whether it answered with the thing users actually need.