Skip to content

HTTP monitors

An HTTP monitor requests a URL on every check and judges the response. It's the right check type for anything users reach through a browser or an HTTP API — websites, web applications, portals, webmail — because it exercises the full chain: DNS resolution, the network path, the web server, and the application behind it.

What counts as up

On each check, UpAll sends a request to the Target URL and applies one rule: any response with a status code below 400 counts as up. That includes every 2xx success and every 3xx redirect. A 4xx (client error, like 404 or 403) or 5xx (server error, like 500 or 502) counts as a failure, and so does no response at all.

Two mechanics around that rule:

  • Redirects are followed, up to 5 of them. If the target redirects — say HTTP to HTTPS, or to a www hostname — UpAll follows the chain and judges the final response. More than 5 redirects means a loop or a misconfiguration, and the check fails.
  • The timeout is fixed at 10 seconds. A target that takes longer than 10 seconds to answer fails the check. This is not configurable — a response that slow is indistinguishable from down for anyone actually using the system. To catch moderately slow responses long before that, use the degraded threshold.

What HTTP monitors don't check

Be aware of the boundary: UpAll judges the status code, not the page content. There are no keyword or content assertions — you can't require that the page contains a certain phrase, and you can't configure an expected status code. A page that returns 200 with an error message in its body still counts as up.

In practice the under-400 rule catches the failures that matter — crashed applications, dead servers, broken upstreams, expired DNS — because those almost always surface as connection errors, timeouts, or 5xx responses. But if a system can fail while still serving 200s, pair the HTTP monitor with a TCP check on its critical dependency, such as its database port.

HTTPS targets get certificate watching

If the target URL starts with https://, the monitor does double duty: alongside the regular checks, UpAll reads the site's TLS certificate about once an hour and opens incidents as expiry approaches or when the certificate is broken. Nothing to configure — see SSL certificate monitoring. This alone is a reason to monitor the HTTPS URL rather than the plain HTTP one.

Practical tips

  • Monitor the URL your users actually visit, redirects included — the final destination is what gets judged.
  • The recorded response time covers the whole request, so the latency chart reflects what a visitor experiences.
  • Before saving, press Test connection to see the live result and response time; see Test connection.