Broken Links vs HTTP Errors

Quick answer

A broken link is a link whose destination cannot be reached as intended; an HTTP error is the response returned by a server for a request. NOTABIS can report both: broken-internal-link describes an internal link and its target status, while http-error describes an affected page response. Fix the link, the destination, or neither only after deciding what the URL is meant to do.

What the issue means

These are related but different observations:

• A page can return 404 when nothing links to it. That is an HTTP error without a broken internal link.
• A link can point to a URL that redirects successfully. It is not broken, but the link may be stale.
• A linked page can return 500, 403, or 429; the link is broken for its intended audience even though the server did respond.
• A URL can show a “not found” message with 200 OK, which is a soft-404 style problem rather than a real HTTP error.

Why it matters for SEO

Broken internal paths waste user journeys and remove a route by which crawlers discover content. Unexpected 4xx and 5xx responses prevent a public page from being retrieved reliably. An intentional 404, 410, 401, or 403 can be correct for a removed or private resource. The status and the link context must be judged together.

How NOTABIS detects it

NOTABIS records internal links from crawled HTML, resolves up to its link-resolution limit, and records broken-internal-link when a target returns no response or a status of 400 or higher. Each crawled page also receives http-error when its final response is 400 or higher. Redirects are recorded separately, and the audit can miss links on uncrawled or authenticated pages.

How to check it yourself

Request the target directly and inspect the source link:

curl -sS -D - -o /dev/null https://example.com/target

Check the final status, redirect chain, authentication, content type, and whether the source link still serves a useful purpose. Search templates, CMS content, navigation, and sitemaps for other references to the same URL.

How to fix it

Restore the intended page and return a successful response, update the link to the final relevant URL, or remove the link when no replacement exists. Use one direct redirect for a genuine move, and let a permanently removed URL return 404 or 410 when appropriate. Do not redirect every error to the home page or return 200 OK for missing content. Re-crawl both source and target after deployment.

When this is intentional or does not need fixing

A removed page, private route, rate-limited endpoint, or temporary maintenance response may correctly return an error. An external link can fail outside your control; label or replace it when it affects users. Do not expose a private target merely to satisfy a crawler.

Related issues

Read broken internal links, HTTP error responses, redirected URLs, and fetch failures.

Run a NOTABIS audit

NOTABIS connects internal link evidence with target response status so you can tell a bad link from a page that is intentionally unavailable.