In one paragraph
Broken internal links means the crawl reached same-site targets that returned error statuses counted by this rule. Identify whether each target moved, was removed, or failed temporarily, then repair the source link or route. Use a relevant permanent redirect for genuinely moved content. Do not send every missing URL to the homepage or treat intentional access restrictions as automatically healthy.
Why it matters
A broken navigation target interrupts a visitor at the moment they expect useful content. A footer error can affect many journeys, while a single product or checkout link can matter disproportionately. Prioritise by actual use and business impact, not just the total number reported.
Crawlers also need working routes to discover content. However, a missing page is not inherently wrong: 404 or 410 can be the correct response for content that has genuinely gone. The problem may be continuing to advertise that target through links or a sitemap.
A server error is different from deletion. Rewriting a 500 route to another page can conceal a real application failure. Diagnose status and intent before selecting a remedy.
How WebAuditOS evaluates this finding
The supplied site-wide result counts crawled targets whose final response after redirects is a 4xx or 5xx, excluding 401, 403, and 429 and recognised bot-filter cases. No counted error passes; one or more fails at medium severity. Timeouts and connection failures are not counted by this rule.
Targets can come from followed links or sitemap entries, so the finding is broader than proof that a visible page contains an offending anchor. The evidence gives an example and target count, not necessarily every source link. Same-host, static-HTML discovery and crawl limits leave other subdomains, file links, runtime-only links, or deeper content outside coverage. Excluded access or rate-limit statuses still need separate review when they affect intended visitors.
How to fix it
- Open the failing URL with an authorised test context and inspect the final status and body. Check whether it is consistently missing, temporarily failing, redirected incorrectly, or blocked for the crawler.
- Locate references in menus, shared templates, page content, and sitemap generation. A stale sitemap can explain a reported target even when no visible link remains. Fix the producer, not just one rendered instance.
- For a typo or outdated source link, update it directly to the correct final URL. Check language prefixes, case, trailing-slash rules, and URL encoding where relevant.
- For genuinely moved equivalent content, add a permanent server-side 301 or 308 redirect from the old URL to its appropriate successor, and update internal links to the final destination. Avoid stacking another hop onto an existing chain.
- For removed content without a meaningful replacement, remove or revise references and return an appropriate 404 or 410. Do not create an irrelevant homepage redirect merely to eliminate an error count.
- For 5xx responses, inspect application and infrastructure diagnostics. Fix the failing route or dependency, preserve meaningful error statuses, and confirm monitoring still detects unexpected failures.
- Review excluded and unavailable cases separately. A 403 can be correct access control or an unintended public-page restriction; a 429 can reflect rate limits. A missing connection response can still be a real availability issue even though this rule does not count it.
- Test source navigation and destination behaviour after deployment, regenerate affected sitemap output, and rescan. Compare coverage so unseen routes are not mistaken for fixed ones.
What the fix can break
Broad redirect patterns can catch legitimate live URLs or create loops. Test representative working paths as well as old ones before deploying a site-wide rule.
Removing navigation without offering a relevant alternative can make a site technically cleaner but less usable. Preserve a useful path when the underlying content still exists.
Caches and bot challenges can obscure the current result. Check the public response and intended audience, not only a successful authenticated browser tab.
Questions
Why was a URL reported when no page links to it?
It may have come from the sitemap or another crawl input. Review those sources before searching only visible content.
Are all 404 pages SEO mistakes?
No. A correct not-found response can be appropriate. Fix references to removed content rather than disguising every missing page as success.
Does an excluded 403 mean the page is fine?
No. It only means this rule does not count it. Verify whether the restriction is intended for the affected audience.
Which related checks help?
Review 4xx and 5xx responses, redirect chains, and sitemaps.