What UptimeCraft does

UptimeCraft watches things and tells you when they break. Everything else is detail, but the detail is worth five minutes because it explains why the numbers say what they say.

The chain

monitor  →  check  →  result  →  incident  →  alert

A monitor is a thing you want watched: a website, a certificate, a DNS record, a nightly job, a server's disk.

A check is one attempt. It runs on a schedule — the interval — from each region you picked. Each attempt produces a result: up or down, how long it took, and what went wrong if anything.

An incident opens when enough checks fail in a row. "Enough" is the failure threshold, and it exists so one dropped packet does not page you.

An alert is the message that goes out when an incident opens, and again when it closes. Where it goes is up to you — see Notification channels.

Why there are regions

A check runs from 10 possible locations: New York, San Francisco, London, Frankfurt, Singapore, Bangalore, Sydney, Toronto, Amsterdam, Atlanta. New York is always on and cannot be turned off; how many others you can add depends on your plan.

More than one region is not about coverage, it is about truth. A site that answers in London and not in Sydney is a site with a routing or CDN problem — real, but different from being down. A single region cannot tell those apart, so it either misses the problem or calls the whole site down.

Failures are counted across all regions together, in the order the results arrive. Three consecutive failures opens an incident whether they came from one region or three. One success closes it.

What you can watch

There are 11 kinds of monitor: HTTP(s), Keyword, Content freshness, Content change, Ping, Port (TCP), Heartbeat, DNS, API, UDP, SSL certificate.

Each answers a different question, and the difference matters — "is it up" and "is it serving the right page" are not the same question, and a monitor that answers the first will happily report a blank error page as healthy. See Choosing a monitor type.

Separately, a server you install our agent on reports CPU, memory and disk from the inside. That is server monitoring, and it works differently: the server pushes to us rather than us asking it.

What it costs you

And how often checks can run, and how long we keep the results:

What we do not do

We check from the outside, like one of your visitors. That means we see what they see, and it also means there are things we cannot see: why it broke, what your logs say, or anything behind a login. We tell you that something is wrong and when it started. Finding out why is still your job — but knowing within a minute, rather than when a customer emails, is most of the battle.