HTTP 410 Gone

404 is an observation: there is no representation for this URL right now. It carries no claim about tomorrow, and a client is entitled to try again.

410 is a statement of intent: this existed, it was removed on purpose, and no forwarding address is being offered. The specification describes it as intended for cases where the resource is intentionally and permanently unavailable.

Google's documentation notes that 404 and 410 are handled in broadly the same way, with 410 leading to the URL being dropped somewhat faster. It is a difference of degree, not of kind.

Neither code is required. A site that answers 404 for everything absent is behaving correctly, and no ranking benefit follows from adopting 410.

Content you deliberately retired and do not want resurfacing: an expired offer, a discontinued product with no successor, a job listing that has closed.

Large deletions after a content audit, where the point is to get several hundred URLs out of the index without waiting for each one to be re-crawled repeatedly.

URLs generated by a bug — a faceted navigation that produced thousands of combinations, for instance. 410 states plainly that none of them will return.

Where a successor exists, neither code is right: a 301 to the replacement serves the visitor and transfers the indexing signals. 410 is for when there is genuinely nothing to point at.

Do not return 410 for something that might come back. Seasonal pages, out-of-stock products and temporarily withdrawn articles are all cases where 404 — or better, a page that stays up and explains the state — is the honest answer.

Do not use 410 as a bulk cleanup for URLs you have not actually reviewed. The code claims deliberate removal, and reversing it means waiting for re-discovery.

Do not pair 410 with a redirect. They are contradictory instructions, and only the first one the server sends is acted on.

Internal links pointing at a 410 are still broken links. The code makes the removal intentional; it does not make a navigation entry that reaches it correct.

`curl -I https://example.com/removed-page` prints the status line. A correctly configured removal answers `HTTP/2 410`.

Most servers and frameworks return 410 from an explicit rule: a list of retired paths in the web server configuration, or a route that responds with the status directly.

In Search Console, 410 responses appear in the Page indexing report under the not-found group alongside 404s, so the two are not separated in reporting even though they are handled slightly differently.

A link checker reports both as broken links when something on your site points at them, which is the finding that actually needs fixing in either case.