Redirect chain

One redirect is not a chain: it is the normal way a moved URL reaches its new address.

Two or more is a chain. `http://example.com/a` → `https://example.com/a` → `https://www.example.com/a` is a chain of two, and it is the most common shape on the web because it is the protocol rule and the host rule firing in sequence.

The length is the number of hops before a 200, not the number of distinct hostnames involved.

Each hop is DNS, connection and response. On a high-latency connection the cost is perceptible.

Crawlers follow a bounded number of hops per attempt. Google documents following a small number in one crawl, treating the chain as unfollowed beyond that — so a long chain can end with the destination never being fetched.

Chains conceal breakage: when the final hop starts answering 404, the first hop still looks like a working redirect.

A loop is a chain that revisits a URL: A → B → A. Browsers stop after a number of attempts and show an error, so the page is unreachable rather than slow.

The usual cause is two rules disagreeing — one adding a trailing slash, another removing it.

A loop is an availability problem and outranks every performance consideration.

`curl -I -L <url>` follows the redirects and prints each hop's status and `Location`. Counting the `HTTP/…` lines gives the chain length.

Test the shapes that produce chains: the other protocol, the other host, with and without a trailing slash, an old path, an uppercase path.

Almost never from one decision. A chain is what accumulates when several correct rules fire in sequence, each added by someone who could not see the others.

The canonical example is a site that added HTTPS, then standardised on the www host, then migrated its URL structure. Each change was right; together they make `http://example.com/old` travel three hops.

Layered infrastructure adds its own. A redirect at the CDN followed by one at the origin is two hops, and the two teams that configured them may never have compared notes.

Trailing-slash normalisation is a frequent third hop, because frameworks apply it after routing rather than before.

This is why chains are worth measuring rather than reasoning about: nobody designed the chain, so nobody can describe it from memory. Following it with a tool is the only way to see what a visitor actually experiences.