Scheduled broken-link and redirect checks that tell you the moment a link breaks — linkbruno.com
You've decided to delete a page. There's no good replacement, so a redirect would just send people somewhere irrelevant. The page is going away.
Now what should the URL return? The HTTP spec gives you two candidates:
- 404 Not Found: the server can't find anything at this URL. It says nothing about whether it ever existed or whether it might come back.
- 410 Gone: the resource used to exist, has been intentionally removed, and isn't coming back. Clients are encouraged to remove links to it.
On paper, 410 is the more precise answer. In practice, very few sites use it. Is that laziness, or are they right? Let's argue about it.
The case for 410
It's the honest answer. The page didn't go missing. You removed it, on purpose. 410 says exactly that, and HTTP has a code for it, so why not use it?
It tells link checkers and crawlers something useful. A 404 might be a typo, a broken deploy or a temporary routing bug. A 410 is a deliberate decision. Tools that see 410 can treat it differently: stop retrying, report "removed" rather than "broken", flag the linking page for a cleanup rather than an urgent fix.
Search engines drop it a little faster. Google has said for years that the difference is small, but 410 is treated as a slightly stronger signal that the URL is gone for good, so it may leave the index sooner. If you're removing a lot of thin or outdated pages, that can matter.
It separates intentional removals from bugs in your own logs. If every deliberate removal returns 410, then every 404 in your logs is something you didn't intend. That makes 404 monitoring far more useful.
The case for plain 404
The practical difference is tiny. Browsers show both as an error page. Search engines treat both as "not here" and stop indexing within a crawl or two. For most sites, nobody will ever notice which one you chose.
410 is easy to get wrong. If you return 410 for a URL by mistake (a bad rewrite rule, a typo in a list), you've told crawlers it's gone forever, more firmly than a 404 would. Recovering from a wrong 410 is slower.
It requires a list. To return 410, the server has to know which URLs used to exist. That means maintaining a list of removed paths forever, next to your redirects. A 404 needs nothing: anything unknown is not found.
Frameworks don't make it easy. Most routers and static hosts have a built-in 404 page and no concept of 410. On static hosting, you might not be able to return a 410 at all without edge functions or server config.
"Never coming back" is rarely certain. Products come back in stock, campaigns repeat, someone recreates the docs page two years later. A 404 doesn't make a promise you might break.
Where I've landed
I use 410, but only in specific situations, and I treat the list of gone URLs like the redirects file: append-only, reviewed, in version control.
- Default to a redirect if there's a page that genuinely answers the same need. Not the homepage; the closest real replacement.
- Use 410 for bulk, deliberate removals where you want crawlers and checkers to stop asking: discontinued product ranges, expired event pages, spammy user-generated pages you've cleaned up.
- Use 404 for everything else, including anything you're not sure about.
- Make both error pages useful. A 410 page can say "this page was removed" and link to the closest section. A 404 page should offer search and the main navigation. Neither should return 200.
- Fix the links on your own site either way. The status code is for the outside world. Your own pages shouldn't link to a 404 or a 410 at all.
Whatever the removed URL returns, the pages that still link to it are the real problem. linkbruno.com checks your site on a schedule and tells you which pages still link to URLs that now return 404 or 410, so a deliberate removal doesn't leave a trail of dead links behind it.
That last point is where most of the real damage happens. The status code argument is mostly about crawlers. The visitor who clicks a dead link in your own navigation doesn't care whether it says 404 or 410, and a scheduled check from linkbruno.com or any link checker you trust is what finds those links.
What a link checker should do with 410
This is the part I'm least sure about. Options:
- Report it as broken, same as 404. It's a dead link from the visitor's point of view.
- Report it separately as "removed". The fix is different: delete or replace the link rather than investigate the target.
- Ignore it on external sites. If another site deliberately removed a page, you can't fix it, and a 410 is arguably the clearest possible signal to just remove your link.
I lean towards the second one, but I can see the argument for treating every dead link the same way.
A quick test for your site
Pick a URL you deleted recently and check it:
curl -s -o /dev/null -w '%{http_code}\n' https://www.example.com/old-page If it returns 200 (a soft 404) or 301 to the homepage, you have a bigger problem than 404 vs 410. If it returns 404 or 410, check whether any of your own pages still link to it. That's usually the bit that's been forgotten.
Related reading
- How to fix 'Failed to load resource: the server responded with a status of 404' (and find which page is asking for it)
- How to fix "Not found (404)" errors in Google Search Console
- Your redirects file should be append-only. Deleting a redirect is a breaking change.
- How to fix 'Page with redirect' in Google Search Console (and when you shouldn't)
So: do you ever return 410? And if you run a link checker, would you want a 410 reported differently from a 404?
