HTTP status codes

The five classes are defined by the IETF in RFC 9110 and catalogued in the IANA registry, so the meaning of a code is the same everywhere. What differs is how a crawler acts on it.

How Googlebot reads each class

Every URL Googlebot touches returns a code, and it responds by class:

• 2xx sends the page into the indexing pipeline, but a 200 does not guarantee indexing. If the page looks empty or shows an error while still returning 200, Google flags it as a soft 404.
• 3xx is followed, up to 10 hops. A 301 is a strong permanent-move signal that consolidates ranking onto the target; a 302 is treated as temporary.
• 4xx pages are not indexed, and previously indexed URLs get dropped. Use 410 Gone to remove content on purpose (Google treats it much like 404).
• 5xx errors make Google slow its crawl and, if they persist, drop the URL. Serve 503 during planned maintenance so crawlers know to return.
• 429 Too Many Requests is the one 4xx that does not remove content. Google reads it as server overload and reduces crawl rate, which is why Google says not to use 403 or 404 for rate limiting.

The wrong code silently leaks rankings. A moved page returning 302 instead of 301 can stall consolidation. A dead page returning 200 wastes crawl budget on a soft 404. A flaky server throwing 500s teaches Google to crawl the site less often.