Check settings and thresholds

Each check has a few settings that decide what counts as a problem. The defaults suit most sites, so you only need to change them when a check is too noisy or too quiet for you.

Where to find them

Open a check and go to its Settings tab. The top of the form shows How often the check runs. That comes from your plan, so there’s nothing to change here. See How often checks run.

Saved settings apply from the next check. If you type a number outside the allowed range, the form tells you the range.

The owner and people who can edit the site can change settings. People who can view see the same values, read-only.

Sitemaps

• Warn when the URL count drops by more than (%): 30% by default. A drop bigger than this, compared with the previous check, is a warning. A drop of more than twice this is critical. If either check couldn’t read every child sitemap, the counts aren’t compared.
• Flag as stale when the newest lastmod is older than (days): 30 days by default. A sitemap whose newest lastmod date is older than this is raised as info.

A sitemap with Google News tags is checked as a news sitemap. Its form hides the two settings above and shows one under News instead:

• Critical when no new article for (hours): 6 hours by default. If the newest article is older than this, it’s critical. Google News only reads articles from the last 48 hours.

robots.txt

• Important paths: none by default. One path per line, starting with /, like /products/. It’s critical if robots.txt stops Google or every crawler from crawling one of them.
• Crawlers to watch: Googlebot, Bingbot, GPTBot, ClaudeBot, PerplexityBot and Google-Extended by default. One user agent per line. You get a warning when robots.txt newly blocks one of them from the whole site. Blocking Googlebot or Bingbot is always critical, even if they’re not on the list. A crawler other than Googlebot or Bingbot that was already blocked on the first check, or when you added it, is taken as deliberate.

ads.txt and app-ads.txt

• Lines that must be present: none by default. One seller line per line: domain, publisher ID, and DIRECT or RESELLER. It’s critical if one of them goes missing from the file. A line that isn’t a valid record is refused when you save.

Uptime and pages

All of these sit under Advanced.

• Expected HTTP status (after redirects): 200 by default. The status the page should end on. Any other status counts as a failed check.
• Slow when a response takes longer than (ms): 5,000 ms by default. The highest you can set stays under the 10-second request timeout, since a slower answer never arrives.
• Failed or slow checks in a row before it counts: 2 by default, so one blip is ignored.
• Text the page must contain: none by default. One per line, not case-sensitive. It’s critical when one goes missing from the page.
• Text the page must never contain: none by default. One per line, like an error message. It’s critical when one shows up.
• Blank when the page has less visible text than (characters): 20 by default. A page with less visible text than this counts as blank, which is a warning. 0 turns the blank check off.

Uptime and the page check explains each of these in more detail.

SSL certificates

• Warn when fewer days are left than: 21 by default.
• Critical when fewer days are left than: 7 by default. It can’t be more than the warning number.

An expired, untrusted or mismatched certificate is critical whatever you set here.

Edit address

Sitemap and uptime checks have their own address. To point one somewhere else, use Edit address in the Address part of its Settings tab.

The check keeps its history, but it starts fresh: the next check is treated as its first, with nothing compared to the old address. If the check is running, the next check runs straight away. The count of failures in a row starts again at zero.

robots.txt, ads.txt, app-ads.txt and SSL checks have no address of their own. They follow the site’s address.

Pause and resume

Pause check on the Settings tab stops a check until you resume it. Its history stays. Any incidents it has open are resolved, and no alerts go out for them.

Resume check starts it again straight away, with its count of failures in a row back at zero. Resume within 10 minutes of pausing, or press Undo, and the incidents the pause resolved open again, with no alerts. A check you paused still counts when you add a new check. To resume it, the site needs a free place: if it already runs as many checks as your plan allows, pause or delete another one first.

Pause site in the site’s Settings does the same for every check on the site. Resume site starts only the checks that were running before.