Content freshness

Fetches a page and works out how old its content is. Goes down when it is older than you allow.

Use it when

Something is supposed to keep changing and failing silently is the real risk: an RSS feed, a generated report, a price list, a status JSON, a nightly export published to a URL.

This catches the failure nothing else does — the file is still there, the server still returns 200, and the contents are three weeks old.

Don't use it when you want to know a page changed; that is the opposite question and the answer is content change.

What to enter

The page itself compares the content to what it was last time and treats a change as freshness — it works everywhere and needs no cooperation from the server.

Last-Modified is exact when the server sends it honestly. Many do not send it at all, and some send the current time on every request, which makes everything look permanently fresh.

A date on the page is for pages that print one. The pattern is a regular expression; lookaround and backreferences are not supported.

Exactly when it's DOWN

• The response status is not 2xx or 3xx.
• No date could be found at all.
• The date found is older than the maximum age.

What it doesn't do

• It cannot tell "the generator is broken" from "there was genuinely nothing to publish". A quiet weekend on a feed that normally updates hourly will alert. Set the maximum age against the longest legitimate quiet period, not the average one.
• No JavaScript, so a date rendered client-side is not visible to it.

Example

URL https://example.com/feed.xml, source "the page itself", maximum age 6 hours. If the feed has not changed in six hours, something upstream has stopped.

Alerts

Down, recovered.