Docs / MonitoringBrowse

Site health

Some breakages never take a site down, so uptime never sees them: a staging noindex that reached production, an analytics tag that quietly went missing, a WordPress plugin with a known hole, a key left in the JavaScript. Site health looks for them once a day.

On this page
  1. What it checks
  2. Verify the domain for the script checks
  3. Alerts
  4. How to fix what it finds
  5. Our crawler

What it checks

Plan/Pro and up

Site health runs on every website monitor (an http or https page opened with GET), once a day at a time of its own, and reads only what any visitor can see. Open the site and go to Health.

CheckWhat it looks at
Search engines
A noindex tag or header, a robots.txt that blocks every crawler, a canonical link to another host.
Analytics tags
Google Analytics, Tag Manager, Meta Pixel, Plausible, Fathom, PostHog, Segment, Hotjar, Clarity, Matomo and Yandex Metrica on the page. Warns when one that was there is gone.
WordPress
The WordPress version, plugins and theme the page shows, against known vulnerabilities and the latest releases.
Secrets in scripts
API keys, tokens and private keys in the page and its own script files. Publishable keys are fine and not reported.
Source maps
Source maps next to the page’s scripts that anyone can download, with your original source code.

A missing analytics tag counts only when a second look, an hour later, still misses it, and only for a tag that was there before: a site that never had analytics is never told about it.

Verify the domain for the script checks

Secrets in scripts and Source maps read the site’s own script files, at most 30 per page. They run only on a domain your workspace has proved it owns.

  1. 1.
    On the site’s Health tab, copy the TXT record. It looks like heedery-verification=….
  2. 2.
    Add it at your DNS provider on the domain shown (example.com covers www.example.com and every other name under it).
  3. 3.
    Press Verify. If the record has not shown up yet, Heedery keeps looking for three days.

NoteOnly the TXT record counts. Keep it: Heedery looks for it once a day, and a domain whose record is gone three days in a row is no longer verified.

Alerts

A new problem goes to the site’s alert email list, the channels that cover the site and a quiet push on the phone. Once per problem, and once more when it is fixed. It never opens an incident, never pages on-call and never sends an SMS or a call: nothing is down.

A check you know about (a private app that is noindex on purpose) can be ignored on the Health tab: it still runs and shows, but never alerts. Run now runs every check again, at most every 10 minutes.

How to fix what it finds

Search engines
Remove the noindex meta tag or X-Robots-Tag header, let crawlers into robots.txt and point the canonical link at this site. These usually come from a staging setting that reached production.
Analytics tags
Check the template or tag manager that loaded the tag, and the consent banner in front of it. If you removed it on purpose, ignore this check.
WordPress
Update WordPress, the plugins and the theme in the dashboard (Dashboard → Updates). Remove plugins you no longer use.
Secrets in scripts
Revoke the key at its provider now and make a new one: it has been public. Then keep secrets on the server, never in code the browser downloads.
Source maps
Stop publishing source maps in production (for Next.js, productionBrowserSourceMaps: false; for Vite, build.sourcemap: false) or upload them to your error tracker only and block *.map on the server.

Our crawler

Site health identifies itself as HeederyHealth/1.0 in the User-Agent, reads at most a few files at a time with short timeouts and a size cap, and never loads scripts from other companies’ servers. WordPress versions are compared with api.wordpress.org and known vulnerabilities come from WPVulnerability.net.

Still stuck? Write to us.