Before an alert goes out, a failed check is re-checked from other places. Even so, some failures are real to the checker and invisible to you. These are the usual ones.
Start with what was seen
Open the incident. Its timeline says what happened in plain words, such as "The site answered with a server error (502 Bad Gateway)." With Link Robins Pro, What we saw adds a screenshot, a likely cause, the headers and how long each step took.
Your host or firewall is turning the checks away
Bot protection, a web application firewall or a host's rate limiter sometimes treats a monitor like an attacker. The checks identify themselves with the User-Agent Link Robins uptime monitor, so a rule that allows that User-Agent is the most reliable fix. The checks come from several places, so allowing by IP address is fragile.
- A 403 or a challenge page means a firewall answered instead of your site.
- A 429 Too Many Requests means "slow down", not "offline", so it never counts as down. If a host keeps limiting the checks, that monitor is checked from a single location instead.
Your site is slow to answer
A request that gets no answer in roughly ten seconds counts as a failure, after the same re-checks as any other. A page that is only occasionally this slow can still trigger it. Pro's Slow-response alert helps you see it coming.
A redirect loop
Redirects are followed, up to five. A loop, or a long chain, ends as a failure ("Too many redirects").
Down in one region only
With Pro you may get "down in that region only". That is not an outage: no incident is opened and your status page does not change. It usually means something in your network treats that region differently, such as a country block or a firewall rule. Check those first.
Short blips
If a brief hiccup is waking you, Pro can wait for several failed checks in a row before alerting. See Link Robins Pro.
Still stuck
Open a ticket from Support, and give the monitor's name and roughly when it happened. That is usually enough to find it.