Broken Links

Linkwise scans the external links across your site, tells you which ones are dead, and lets you fix them in place — without opening each entry.

What it does

It finds external URLs that no longer work — 404s, expired domains, certificate failures, timeouts — and groups them so you can replace, ignore, or remove them in a few clicks.

How it works

For every external link Linkwise makes an HTTP request and classifies the result:

• A fast HEAD request first; if the server rejects HEAD (many WAFs and CDNs do), it confirms with a GET before flagging anything — so HEAD-hostile hosts aren't reported as broken by mistake.
• Failures are classified as not found (404 and similar client errors), forbidden (401/403), server error (5xx), timeout, SSL error, or connection/DNS failure.
• Transient failures are retried (twice by default). A momentary timeout or refused connection is not cached as broken — the next scan re-checks it, so a blip doesn't pin a working link as dead. Authoritative results (a real 404) are cached against the entry's content hash so re-scans stay fast.
• The first-detected date is preserved across scans, so you can see how long a link has been broken.

Using it in the Control Panel

1. Open Linkwise → Broken Links and click Check Links. The scan runs in the background and is cancellable on large sites.
2. Review the results, grouped by entry. Each row shows the URL, the error type, and where it appears.
3. For each broken link, choose inline:
   • Replace — enter a new URL; Linkwise rewrites the link in place.
   • Ignore — mark a false positive so it stops appearing.
   • Unlink — remove the link, keeping the anchor text.
4. Select several and bulk-unlink them in one operation. Export the report as CSV. On multilingual sites, filter by locale.

Settings

Option	Default	Effect
broken_links.timeout	10	Seconds to wait per request.
broken_links.retries	2	Retries before declaring a transient failure broken.
ignored_links	''	Newline-separated URL patterns to skip (never reported).

ignored_links patterns match the full URL and support * as a wildcard. To ignore every link to a domain, use *example.com* — a bare example.com won't match https://example.com/page.

Notes & limits

• Only external links are checked here. Internal links pointing at missing/unpublished entries surface through the index, not this scan.
• A flaky target can produce a transient error one scan and clear the next — that's by design; re-run Check Links if a result looks wrong.
• Replacing a URL here changes the link in your content; the change is recorded in the Activity Log.