A page check opens one page in a real browser and confirms something is true about what came back. It does not click and it does not type. It is the monitor to reach for when the question is "is this page still right", not "does this journey still work".
What it does
- Opens the address you gave it in a real browser.
- Waits for the page to settle.
- Checks the one thing you said should be true.
Because it runs a real browser rather than fetching the page, it sees the result after JavaScript has run. A page that returns a normal response and then renders an error still fails this check.
What you can assert
- Text is present on the page.
- Text is absent, which is how you catch an error message appearing.
- The title contains something.
- The address matches a pattern, which catches an unexpected redirect.
- An element exists, chosen with the visual picker or named yourself.
- An element does not exist.
- At least N elements exist, which is how you catch a listing that came back empty.
Good uses
- A product page still shows its price and its buy button.
- A collection or listing page is not empty.
- A sign-in page loads and its password field is present, so a deploy that breaks the form is caught.
- A key landing page has not started showing an error.
What it will not catch
A page check proves the page rendered. It does not prove anything works. A product page can pass while the add-to-cart button underneath it does nothing, because a page check never clicks it. For that, use a guided click check.
Setting one up
From Monitors, choose Create monitor and pick a page-check template. Give it the address, then choose the assertion and its value.
Start with the assertion that would actually be wrong if the page broke. "Text is present" on a word that only appears when the page is healthy beats checking for the site name, which will be there whatever happens.