Link previews and search
What a shared link shows, the tab icon, how search engines and AI crawlers see a status page, and what the searchIndexing setting changes.
There is nothing to set up for any of this. It all comes from the page's name, description, logo and settings.
Link previews
A shared link to a page shows a card when the site it's shared on supports previews. The card has:
- your logo, if it can go on a card
- the words "Status page", in English whatever the visitor's language
- the page's name and its description, shortened to fit
- the page's host
- a small Norcube wordmark
A card never shows the current status. Social networks keep a preview for days, so a card saying "All systems operational" could still be on display during an outage. The preview's title and description stay away from the status for the same reason.
Renaming the page or changing its description or logo gives it a new card image. An incident doesn't. Networks that have already stored the old card may keep showing it for a while.
When you get the plain card
The card is drawn with fonts that cover Latin text, accents included. A name or description with characters they lack, such as Chinese, Cyrillic or Greek text or an emoji, gets the plain Norcube card instead, because those characters would otherwise draw as empty boxes. The plain card also stands in if drawing a page's card fails.
A logo that can't go on a card leaves the rest of the card in place. That applies to WebP images and to SVGs with elements beyond simple shapes.
Preview links and error pages have no card. A page kept out of search still has one, since sharing a link is your choice.
Tab icon
The icon in the browser tab follows the page's verdict: green, amber or red, the same as the status shown on the page. It's grey when there's no current verdict, such as while updates are delayed.
The icon updates when the page is reloaded. An open tab doesn't change on its own.
The 404 page and the Apple touch icon, which people get on a home screen, use the Norcube mark, because neither has a status to show.
"Last updated" and "Checked"
The times on a page are the last heartbeat from Norcube Monitor, so a quiet page doesn't look stale. If the heartbeat is missing, or the content changed after it, they show the time of the last content change.
A copy served during an outage keeps the time it was built with, and never claims
a newer one. asOf in page.json
and the feed always means the last content change.
Search indexing
searchIndexing is one of a page's required settings, so every page has made the
choice. It decides how the page presents itself to crawlers:
With searchIndexing | true | false |
|---|---|---|
| Page | Can be indexed | noindex, in a meta tag and a header |
| Canonical and hreflang | Present | None |
| Structured data | On the home page | None |
sitemap.xml | Listed | 404 |
llms.txt | Served | 404 |
robots.txt | Allows everything, names sitemap | Allows everything, names no sitemap |
robots.txt keeps allowing crawlers when the setting is off, because a crawler
has to be able to read a page to see its noindex. Turning it off asks search
engines to leave the page out. It doesn't hide the page: anyone with the address
can still open it, and page.json and the feed stay available.
The current status stays out of titles, descriptions and tags, because search results keep them for days after it changes.
What each host serves
Each of these belongs to the host it's served on:
| File | What's in it |
|---|---|
/robots.txt | Allows everything, with no rules for particular bots, and names the sitemap of the address that gets credit. |
/sitemap.xml | The home page, the history and each incident, with the dates the page really holds. |
/llms.txt | For AI assistants: the name, description and links to page.json, the feed and the history. No status. |
The home page also carries JSON-LD structured data: WebSite and Organization,
with your logo's address when you have one. The page's name stands in for the
organization's name. Incident pages have none.
These responses carry X-Robots-Tag: noindex, as do page.json and the feed, so
the files themselves don't show up in results. Logo files and the card image don't.
Custom domains
Once a page has an active custom domain, search engines are told to credit that domain:
- The canonical address on both hosts points to the custom domain.
- The default address keeps working with no redirect.
- The default address's
robots.txtnames no sitemap and itssitemap.xmlis a 404, so it doesn't compete with the custom domain. canonicalUrlinpage.jsonholds the custom domain's address.
Each language's ?lang= address is its own canonical address. Other query
parameters are dropped from it.