Redirect tracing: every hop, including hidden ones

The three kinds of hop

A chain can mix all three kinds, which is why reading one part tells you little.

• Server redirects

  The response says the thing lives elsewhere, and the client goes there. These are the ordinary case, they are fast, and every tool sees them.

• Meta refresh

  The page loads, and inside it is an instruction to go somewhere else after a moment. A client that does not process page content never follows it, and one that does pays for loading a page it will immediately abandon.

• Redirects performed by scripts

  The page loads, code runs, and the location changes. These are invisible to anything that does not execute scripts, which includes a great deal of automated traffic.

The distinction matters because a chain that looks like one hop from the server's point of view can be three from a visitor's, and the last two kinds cost a page load each.

Why the requesting client matters

The same address can produce different chains for different callers, and this is not exotic behaviour.

Sites redirect by device, by language, by region, and sometimes by whether the caller looks automated. A trace made as one kind of client therefore describes that client's experience rather than the address, which is why being able to repeat the check as a different one turns a single observation into a comparison.

Where the chains differ, the difference is the finding. A path that sends crawlers somewhere other than where it sends people is worth knowing about deliberately rather than discovering later.

What the trace tells you

Three things, and the third is the one people come for.

• Where you actually end up

  The final address, which is frequently not the one anybody intended after a few years of accumulated rules.

• How many hops it took

  Each one is a round trip that produced no content, and chains grow by accumulation rather than by design, as covered under redirects.

• Which rule seems to be responsible

  Not the rule itself, which lives in your configuration, but the shape of each hop: what changed between one address and the next. A prefix appearing, a slash being added, a parameter being dropped, each points at a different rule in a different place.

What a trace does not report

One boundary, stated because trace tools are often expected to grade things.

A trace does not tell you whether a redirect is correct, whether search engines approve of it, or what it does to your rankings. Those are questions about intent and about engine behaviour, and the second changes without notice. What a trace reports is what happened when it asked, which is the part that can be verified.

FAQ