Canonical URLs Are Hints, Not Redirects
One page can appear under several URLs because of tracking parameters, filters, print views, or site configuration. A canonical link tells search engines which version you would prefer them to treat as the main one.
Add it to the document’s <head>:
<link rel="canonical" href="https://www.example.com/product1">
If these URLs show the same product, they can all point to that preferred version:
https://www.example.com/product1
https://www.example.com/product1?color=blue
https://www.example.com/product1?size=large
A canonical does not send the visitor anywhere
Unlike an HTTP redirect, the canonical link does not navigate the browser. People can still open the alternative URL. It is a signal for search engines, and they may choose a different canonical when other signals disagree.
That is why canonical tags should agree with internal links, sitemaps, redirects, and the actual content. Use an absolute URL, make sure the preferred page is indexable, and avoid chains where one canonical points to a page that canonicalises somewhere else.
Duplicate content is not automatically a penalty. The practical problems are diluted signals, wasted crawling, and search engines choosing a version you did not want in the results.
Use a redirect when an old URL should no longer be accessible. Use a canonical when multiple useful URLs need to remain available but represent the same or substantially similar content. Those jobs overlap slightly, but they are not interchangeable.