Crawl capacity and crawl demand
Google combines two ideas. The crawl capacity limit is how many parallel connections and how much server time Googlebot can use without overloading your site: it rises when the server responds quickly and drops after 5xx errors or 429 rate-limit responses. Crawl demand is how much Google wants to visit your pages: it depends on how many URLs it knows, how popular they are and how long since its last visit. A site migration raises demand temporarily.
Who should actually care
Google points to three profiles: large sites with around a million unique pages that change weekly, sites with around 10,000 pages that change daily, and sites with many URLs stuck as "Discovered, currently not indexed" in Search Console. A 40-page business website whose new pages get crawled the same day has no crawl budget problem.
The most common crawl waste
- Faceted navigation creating thousands of filter URLs: block those parameters in robots.txt.
- Duplicate content that was never consolidated.
- Deleted pages returning 200 (soft 404s) instead of 404 or 410.
- Long redirect chains.
- Noindexed pages: Google still has to fetch them, so noindex does not save budget.
- Sitemaps listing stale URLs or without a trustworthy lastmod.
How to measure crawling
The Crawl stats report, under Settings in Search Console, shows requests per day, average response time and breakdowns by status code and file type. For deeper analysis, server logs reveal which sections Googlebot actually visits: it is the only way to know whether your product pages get crawled or the bot burns its visits on sort orders. A fast server that supports 304 Not Modified responses also lets Google reuse its cached copies.
