Hreflang alternate explicitly requests no indexing

Why it matters

The alternate relationship points to an address whose own directive excludes indexing.

What to do

1. Inspect the captured X-Robots-Tag and HTML robots metadata on the alternate.
2. Remove noindex when accidental; otherwise remove the address from the intended indexed alternate set.
3. Recheck both the target response and the original declaring pages.

What we read

The audit reuses this run’s admitted page and canonical-target captures, then selects at most 20 additional distinct HTTPS alternates across internal and external hosts. Requests use the shared robots, safety, pacing and redirect controls. A probe is not a billed page. The captured terminal document is parsed once; JavaScript and foreign sitemaps are not read.

Only a 2xx answer at the declared alternate qualifies. An explicit X-Robots-Tag noindex or HTML meta noindex fires; the header is also read on non-HTML media.

A clean result requires a known header directive scope and complete applicable meta-robots observation. Missing or incomplete absence evidence remains unknown.

An alternate that redirects is not applicable here because the terminal page’s directive is not the redirecting address’s own directive. Non-2xx responses belong to the status check.

Example & evidence

Illustrative example — declared addresses and responses, not a live audit.

A declares B as an alternate. B returns 200 with X-Robots-Tag: noindex.

Correction: remove the unintended directive or choose the intended indexable alternate.

Confirm & check again

Check for yourself: Check physical X-Robots-Tag fields and the captured robots metadata, rather than assuming the absence of one visible tag proves indexability.

Check again: Full audit — Run another audit of the same declared target and its declaring pages. A sufficient clean check can fix the same target finding. If the declaration disappears, every previously affected page needs complete HTML, HTTP Link and own-sitemap coverage, a well-formed target set and agreeing admitted aliases. Missing, malformed or unread evidence carries the finding.

Limits & sources

The check establishes an explicit directive, not all conditions for indexing. The page-level noindex view and the alternate relationship describe different facts and can both remain visible.

A refused, over-budget, failed or unread target remains unknown. Different or unread admitted aliases qualify declaration completeness; their declarations are never merged into the representative page. These observations do not establish Google’s indexing or chosen canonical.

Confidence: High High for the retained response or declarations that directly establish this defect. Coverage gaps remain separate and never count as a clean result.

Sources

• Google, Block search indexing with noindex — Explains the noindex directive in HTML metadata and HTTP headers. Checked 3 November 2026.

Editorial review: 3 November 2026