Add a self-referencing rel=canonical on every indexable page pointing at the preferred HTTPS URL. That tells crawlers which version to consolidate when parameters, mirrors, or slash variants exist.
SEOHub reports missing canonicals, invalid canonical URLs, multiple canonical tags on one page, and site-wide coverage gaps when many crawled pages omit the tag. Cross-domain canonicals that point off your host are called out separately.
Use canonicals for query-param and trailing-slash variants — don’t noindex pages you still want indexed. Canonicals consolidate; noindex removes.
Prefer absolute https:// URLs in the href. Relative or http:// canonicals are a common source of “invalid canonical” flags after a protocol migration.
One tag only. Multiple conflicting canonicals confuse parsers and often appear when a theme and a plugin both inject link tags.
Audit redirect behavior with the redirect checker so the preferred URL actually resolves in one hop — a canonical that 404s or chains through three redirects wastes the signal.
Confirm robots rules with the robots.txt inspector before you rely on canonicals alone. A Disallow that blocks the preferred URL cannot be fixed by a tag on a crawlable duplicate.
After fixing, crawl again and confirm “No canonical” / coverage flags disappear in the site map and checklist panels.
For pagination, faceted nav, and print views, decide a policy: self-canonical each useful view, or point filters back to a clean category URL. Document the choice so new templates do not invent a third rule.
Watch www vs non-www and trailing-slash hosts. Pick one hostname, 301 the rest, and make canonicals match the live preferred host.
When comparing competitors, note whether they canonicalize category filters — inconsistent handling of facets is a frequent crawl-budget leak on large catalogs.
Ship canonicals in the HTML head of the first response when possible. Client-only injection is slower for crawlers and easier to miss in static audits.