Canonical points to an error page
What this means
The rel=canonical target returns a 4xx or 5xx. The page is nominating a URL that does not exist as its indexable identity, so the canonical signal is discarded and the page risks being dropped or consolidated against an unrelated URL.
How to fix it
Repoint the canonical at a working 200 URL — usually the page's own address — or restore the target URL if it was removed by mistake.
Fix effort
This check needs review: the change can be drafted for you, but a person has to approve the wording or choose the target before it ships.
Authoritative source
https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls
Find every page this affects
CrawlX runs this check — and 161 others — across your whole site, ranks the findings by estimated impact, and opens the fixable ones as pull requests. Free for 500 URLs.