Redirect checker

Requests the URL and records the status: 301, 302, 307, 308 or a final code.

Reads the `Location` header and repeats, keeping the order of hops.

Stops on a final response, on a loop, or at a hop limit.

Reports the chain: each URL, each status, and the final destination.

Protocol and host variants converging on one canonical address in at most one hop.

Permanent moves using 301 or 308, temporary ones using 302 or 307 — and the choice matching reality.

The path preserved: `http://example.com/a/b` reaching `https://example.com/a/b`, not the home page.

A final response of 200.

Three or more hops, typically the accumulation of a protocol rule, a host rule and a path rule that were never collapsed.

A 302 on a move that happened years ago, which keeps the old URL indexed.

A chain ending in 404: the first hop looks like a working redirect and the destination does not exist.

A loop, where the chain returns to a URL already visited. Browsers give up and the page is simply unreachable.

The four host and protocol variants: http and https, apex and www.

A deep path, not only the home page — rules covering `/` alone are common.

With and without a trailing slash.

An old path you know was moved, and an uppercase variant of a working path.

`transport.redirect_chain` records the hops for the URLs encountered during the crawl. `transport.http_redirect` checks that plain HTTP reaches HTTPS, and `transport.canonical_host` that the host variants converge.

`availability.redirect_loop` reports loops separately: a loop is an availability failure, not a performance one.

Read it in three passes, because a chain carries three different kinds of information and mixing them is how the wrong hop gets edited.

The destination first. If the final response is not 200, nothing else about the chain matters — a chain ending in 404 or 500 is a dead URL wearing a redirect, and the intermediate hops are a distraction.

Then the status codes, hop by hop. A 302 anywhere in a permanent move is the finding: it is telling every search engine to keep the source URL. A 307 or 308 where you expected 301 or 302 is usually correct and simply a newer configuration.

Then the path, hop by hop. `/blog/article` reaching `/` means a rule matched too broadly and discarded the path, which is a worse defect than any number of extra hops — every old URL lands on the home page and the mapping is gone.

Only then the length. Three hops that end in the right place with the right codes is a performance note; one hop that lands on the wrong page is a content failure.

Flattening means pointing the first URL directly at the final destination, not deleting the intermediate rules — those still serve the other URLs that enter the chain partway along.

Where the hops come from different layers — a CDN rule then a server rule — the earliest layer is usually the one to edit, because it can answer without the request travelling further.

Internal links are the cheaper fix and are often the real cause of the traffic: a redirect exists for links you do not control, and every internal link pointing at a redirect rather than its destination is a hop the site imposes on itself.