Guide
Getting started
The first quarter of an hour: an account, a server reporting, a site watched from outside, and somewhere for an alert to land.
Step by step
- Open an organization with Start free. No card, and the free plan is enough to see the whole product: 3 servers · 3 website monitors · 7 days of history.
- The first screen says Nothing is monitored yet and hands you an install command with your own token already in it. Paste it on a server. It reports in under a minute.
- Look under Hosts. The server is there with its checks already found — disks, memory, services, containers, certificates — and every one of them already carries a rule that waits before it says anything.
- Go to Website monitoring and press + New monitor. Give it the address your users type. It is loaded from our points of presence, from outside your network, which is the half a server cannot tell you about.
- Go to Alerting. Create one channel under Channels, then one policy under Policies so an incident reaches it, and press Try it to see who would be told. Nothing is sent.
- Incidents is the screen you will live on. Panel is the numbers and Events the history; the Connect group in the sidebar is how a script or an assistant does all of the above without you.
What you have after that
Two kinds of watching, meeting in one place. The agent on the server sees the disk filling and the service that died; the monitor outside sees what a customer sees. When both go wrong at once you get one incident, not two pages from two products.
None of it waits on us: thirty days of Team, then a free plan that stays. When you outgrow it the portal says so, in the screen where the limit bit.
The one thing people get wrong
Adding servers and stopping there. A fleet with no channel and no policy monitors beautifully and tells nobody — the incident is raised, correctly, on a screen nobody is looking at at four in the morning. One channel and one catch-all policy on day one; narrow them later, when you know what actually wakes you.
Where to go next
The agent in detail, and then where an alert actually goes.