Capture the response chain first. Interpret status, headers, and HTML as separate signals before combining them.
Record request context
Write down the exact URL, method, timestamp, and relevant request headers. Scheme, host, path, query, cookies, authentication, language, and user agent can change the response.
Read the chain before the destination
For each hop, record status and Location exactly as returned. Resolve relative locations against the current URL, note loops, and stop at a defined hop limit.
The final 200 does not erase a problematic chain. A temporary hop, cross-host change, protocol downgrade, or loop belongs in the diagnosis.
Keep header and HTML directives separate
A robots meta element lives in HTML. X-Robots-Tag lives in the response headers and can also control non-HTML resources. Record each occurrence and the user agent it targets.
A canonical link is not a redirect and a redirect is not a canonical declaration. They can point to the same destination, conflict, or be absent.
If you only paste the final HTML into an inspector, the original status, Location header, and X-Robots-Tag are unknown. Say so in the result.
Translate evidence into the next check
Use the observed response to choose a focused follow-up: inspect the server rule, compare an unauthenticated request, fetch the canonical target, check an alternate host, or verify the rendered document.
Do not rewrite production redirects or indexing controls from one capture when the route varies by region, device, cache, authentication, or deployment.