Watchtower documentation
Watchtower watches over your network. You install a small agent on the machines you want to keep an eye on, and the console shows you what they are doing — live.
Start here
| If you want to… | Read |
|---|---|
| Understand what Watchtower does before installing anything | What is Watchtower? |
| Get a first machine reporting | Quickstart |
| Install the agent properly | Install the agent |
| Add machines without pasting keys around | Enrolling machines |
| Work out why a machine is not showing up | Troubleshooting |
How the pieces fit together
Three things, and only one of them runs on your hardware:
- The agent — a small native program on each monitored machine. It dials out to Watchtower and reports what it finds. Nothing ever connects in to your machines, so you do not open firewall ports or forward anything.
- The broker — the service the agents connect to, and the same one the console connects to. It moves telemetry between them.
- The console — the web interface at watchtower.page, where you actually look at your fleet.
Machines only stream detailed metrics while you are looking at them. When a machine's card scrolls out of view, its agent stops sending until you scroll back. This keeps the traffic proportional to what you are actually watching, and it means a large fleet does not cost you a large amount of bandwidth to own.
One thing to know up front
Some hardware is not collected yet. The agent reports operating system, CPU, memory, disks, uptime, and logon sessions from the real host. GPU, WAN IP, geolocation, and network adapters report as empty rather than guessing — that is honest absence, not a failure, and nothing you configure will populate them. See What the agent collects.
Getting help
Email support@watchtower.page, or see the support page.