You canonicalized your parameters and think the crawl problem is solved

You canonicalized your parameters and think the crawl problem is solved

The assumption: point every ?utm= and ?sort= URL's canonical at the clean version, problem fixed.

Actually, canonical is an indexing hint, not a crawl directive. Google still crawls every parameter URL to read the canonical tag. It still spends the request. On a site with combinatorial parameters, that's the entire crawl-budget leak — and canonical does nothing to plug it.

Google is explicit: canonical consolidates indexing signals, it does not prevent crawling.

The fix: for genuinely useless parameters (tracking, session), disallow them in robots.txt so the crawl never happens. Reserve canonical for parameters you still want crawled but not separately indexed.

Canonical cleans the index. It does not clean the crawl. Different jobs.