Revalidation and the invalid status

An item is not checked once and trusted forever. The web behind it moves: posts are unpublished, addresses change, hosting starts turning robots away. Traffman keeps checking, and that repeated checking is revalidation.

Two things trigger a repeat check

• The scheduled cycle. Items already marked invalid are retried on their own, on a schedule that slows down the longer an item keeps failing.
• A click. When a visitor clicks your content in a block on another member's site, the service verifies the item at that very moment, before sending the visitor on. The result is remembered for a while, so a burst of clicks costs your server one visit rather than one per visitor.

Both paths run the same checks – see how validation works – and both update the state stored for the item.

Why an item turns invalid

• The page is gone or no longer public. Deleted, switched back to draft, made private or password-protected.
• The address changed. The service still holds the old one, and only a new push replaces it.
• The server was slow or unreachable. A page that keeps the check waiting counts as a failure, even though a patient reader would have seen it.
• A firewall, CDN or bot filter answered instead of the page. Often a hosting option nobody switched on deliberately – see when hosting protection blocks our server.
• A consent screen, paywall or age gate is served to a visitor without cookies. The check receives the wall, not the article, so the text no longer matches.
• The page is built by scripts. There is no article in what the server sends, only an empty frame.
• The text was rewritten. Once the page has little in common with what you offered, the item stops matching it.

The first three come and go on their own, which is exactly what the retry cycle is for. The rest need something changed on your side, or a fresh push.

What a failed check costs you

If the check fails at the moment of a click, the visitor lands on our content unavailable page instead of your site. The click is not counted: nobody earns points for it and you spend none, but the traffic is lost. That is why an invalid item is worth fixing rather than waiting out.

How repeat checks are spaced out

The service comes back to invalid items on its own, but neither at a constant rate nor forever. The first repeat follows within hours; after each further failure the wait grows – to days, then to weeks. A site that was briefly down is therefore picked up quickly, while an item that has been broken for a long time is not requested over and over.

After a long run of failures – around a month and a half – the item leaves the cycle and is not retried automatically again. A single successful check at any point resets all of it: the item becomes valid, and any later failure starts counting from the beginning.

Content of an inactive client is skipped entirely: it is not offered to anyone, so there is nothing to recheck.