The real page is somewhere else

This page loads, renders, and exists. Its canonical, though, asks search engines to index another URL instead.

What pointing elsewhere instructs

A rel=canonical that names another address is an instruction, not a suggestion: index that URL as the real page and route the ranking signals there. It earns its place on syndicated copies, parameter variants, and print versions of articles. Set deliberately, it settles the duplicate-content question before anyone asks it.

When it stops being deliberate

The damage starts with a template mistake. Hardcode the homepage into every canonical, or worse a staging domain, and each page on the site tells search engines to defer to one URL instead of standing on its own. Sections leave the search results without a single error appearing anywhere. That is why an audit flags every canonical pointing away from the page that carries it.