Search Console's "Sitemap could not be read" is one message for four different failures. That is why the usual advice — resubmit it, wait, resubmit again — works for maybe a quarter of the people who try it.

The message means Google fetched the URL you submitted and could not parse it. Nothing about your pages. The four causes below account for every case we have seen across six small static sites, and the first command below tells you which one you have before you change anything.

curl -sS -o sitemap.xml -w '%{http_code} %{content_type}\n' https://example.com/sitemap-index.xml head -c 300 sitemap.xml; echo 

One: it is HTML, not XML

The most common cause on hosts that rewrite URLs or gate traffic. You request /sitemap.xml, the edge serves your 404 page or a browser challenge, and the body starts with <!doctype html>. Google reports it as unreadable because an HTML document is not a sitemap.

The content_type in the first line of the output above settles it. text/html on a sitemap URL is the failure; application/xml is not.

Two: the index points at a child that is gone

A sitemapindex is a list of other sitemaps. If the child file was renamed by a build — Astro and other generators name them sitemap-0.xml until the page count changes — the index keeps pointing at a name that no longer exists, and Google reports the index as unreadable.

curl -s https://example.com/sitemap-index.xml | grep -o '<loc>[^<]*' | sed 's/<loc>//' 

Fetch each URL that prints. Any 404 there is your cause.

Three: it declares another host

A sitemap has to list URLs on the host it is served from. Preview deployments are where this bites: a pages.dev build writes its own sitemap, someone copies the file to the custom domain, and every <loc> names the preview host. Google will read the file and reject the contents.

curl -s https://example.com/sitemap-0.xml | grep -o '<loc>[^<]*' | head -3 

If those lines do not start with your domain, that is the whole problem and no resubmission fixes it.

Four: it is too big, or compressed in a way that breaks it

The published limits are 50,000 URLs and 50 MB uncompressed per file. Small sites hit this only by accident — usually a generator that emits the whole catalogue into one file after a bulk import. Gzip is fine as long as the file is served with Content-Encoding: gzip; a .gz file served as application/octet-stream is not.

What to do after you find it

Fix the file, deploy, then resubmit once. Not five times. Each submission resets the clock on the same check, and you lose the ability to tell which change fixed it.

There is one thing worth doing in parallel, and it is not on this list: confirm the sitemap is not the only path Google has to your pages. A sitemap is a hint. Internal links are the actual map — a URL that appears only in a sitemap can sit in "Discovered" for weeks, while the same URL linked from two indexed pages gets fetched in days.

We hit this shape on our own site in the first week: the index was live and submitted, its child 404'd after a config change, and the error read like a content problem for two days. The version of this checklist we run now, with the exact command per row, is at sitemap could not be read, and the surrounding pipeline — every Page indexing row in crawl order — starts at Google indexing issues.