Selected answers from the Dumb SEO Questions G+ community.
Michael Martinez: If the site is generated by a Content Management System (CMS) like WordPress, Joomla, or Drupal then just grab a couple of pages at random and look at their canonical declarations. You don`t need to crawl an entire site. SEOs waste a lot of time and resources on unnecessary site crawls.
Alan Bleiweiss: While looking at a sampling is smart, the larger the site, even when using a CMS, the greater the potential for missing problems because rules grow exponentially over time and are not template level implemented.
So if it is a large site, it is wise to crawl with a tool that reports canonical issues, such as when meta robots are set to noindex yet have self pointing canonicals.