Your dead URLs that still receive links

The ones marked “gone”, top of the list first. Those URLs genuinely no longer exist, so every link pointing at them is wasted and every visitor following one hits an error page. A 301 redirect to the closest living page recovers the value immediately, without asking anyone for anything.

Because “broken” here means the URL did not answer with a 2xx when it was checked. A page behind bot protection answers 403 to the crawler and 200 to your visitors; a page on a server that timed out answers 521 or 522 for a few minutes and then works again. Both look identical to a link crawler, which is exactly why the status code is shown here rather than hidden behind one count.

Not to the page. It is worth checking that your bot protection is not also turning away Google and the AI crawlers, because the same rule usually applies to all of them. But the page itself is fine and needs no redirect.

Both come from Cloudflare, not from your site. “Refused” (521) means Cloudflare reached your server and the connection was actively rejected — nothing listening on the port, or a firewall dropping Cloudflare. “Timeout” (522) means the connection was never answered at all within about fifteen seconds — the server was overloaded, too slow, or silently discarding the traffic. Refused points at configuration, timeout points at capacity.

It is a hosting question, not a link question. A handful across different dates is normal noise. A large block of them, all recent, means your server was down or throttling when the crawl ran — worth a look at your uptime, not at your redirects.

To the closest living equivalent — the same product, the same article, the same category. The anchor text column is the best clue you have about what the linking site was pointing at. Send it to the home page only when nothing better exists: a home-page redirect keeps some authority but throws away all the relevance.

The count in the backlink profile comes from the whole-profile summary and includes historical and lost links; this table is built from the live broken links themselves, grouped by address, so it holds the ones a redirect would actually recover today. When the two disagree by a lot, trust this table — it is the one made of individual URLs you can open.

It is the last data collected for your site — nothing is fetched live when you open this page. Each block of data is re-collected on its own schedule by the automatic tasks. The report page shows the collection status of every section, and the Re-collect all data button there starts a fresh round when you have applied your fixes.

It does not — the report simply shows the first 25 rows as a preview, and this page shows every row that was stored. What can differ is the number quoted in a check point: those are often counted over the whole domain by DataForSEO, while this table only holds the rows that were downloaded and kept.

Click any column heading to sort by it; click it again to reverse the order. Rows with no value stay at the bottom either way. Pages hold 100 rows and keep your sort as you move between them, so the link you copy always reopens the same view.

Use it to pick targets, not to admire numbers. Sort by the column that matches your goal, take the first handful of rows, and turn them into concrete work — a page to write, a link to ask for, a title to rewrite. The report's action plan already ranks the most urgent of those for you.