Canonical URL With a Query String: When It's Fine and When to Fix It

What this warning means

The page's canonical URL contains a ? query string. It's a review item rather than an automatic error, because some parameterized canonicals are intentional.

When a query string in the canonical is correct

• Paginated series: page 2 should canonicalize to itself (/blog?page=2), not to page 1. Otherwise the posts linked only from page 2 lose a crawl path.
• Sites whose real page URLs use parameters, for example /product?id=4812, where the parameter identifies the content.
• Filtered or faceted pages you deliberately want indexed as landing pages, such as /shoes?color=red.

When it's a bug

• Tracking parameters: utm_source, gclid, fbclid, ref.
• Session or user identifiers: sessionid, sid.
• Presentation-only parameters: sort, view=grid, currency.

The usual cause is a template that builds the canonical from the full request URL, so whatever parameters a visitor arrived with get copied into the tag:

<!-- request: /pricing?utm_source=newsletter -->
<link rel="canonical" href="https://example.com/pricing?utm_source=newsletter">

How to fix it

Build the canonical from the route or the stored page URL, not the raw request, and keep only parameters that change the content:

<link rel="canonical" href="https://example.com/pricing">

In a server framework this usually means using the request path instead of the full URL. In Flask, for example, use request.base_url or url_for(...) with _external=True instead of request.url. After deploying, crawl the site again with a few tracked URLs as start points and confirm the canonical no longer changes.