Server monitoring works from the inside. A small agent on your server reports CPU, memory, disk and network usage every minute over HTTPS. It runs on Linux, macOS (Intel and Apple silicon) and Windows, and never reads files or logs.
Create the server monitor
- Open Monitors, choose the Servers filter and then New server.
- Give it a Name.
- Set the limits: CPU alert at (%), Memory alert at (%) and Disk alert at (%), and For at least how long, so short spikes don’t alert. Leave a limit empty to turn it off.
- Choose Down after no data for: how long the agent may be silent before the server counts as down.
- Keep Alert when the server restarts on to hear about reboots.
- Tick the alert channels and choose Create monitor.
Install the agent
The server’s page opens with Install the agent and a one-line command. Copy it and run it on the server:
- Linux and macOS: in a terminal, with sudo. It works on x86-64 and ARM: Ubuntu, Debian, RHEL, Alma, Rocky, Amazon Linux, Alpine and macOS. The download is checked against its SHA-256 checksum.
- Windows: in PowerShell as Administrator. It installs a Windows service named Uptimecheck Agent, on Windows Server 2016 or later (x86-64).
Turn on Update the agent automatically before you copy the command, and the agent installs new releases signed by Uptimecheck by itself. Without it, the server’s page tells you when an update is available. The charts fill in once the first report arrives.
Good to know
- The command contains a token that belongs to this server. Add a new server monitor for each other server.
- Every disk is listed, and the fullest one is charted.
- To check, update or remove the agent later, see server agent commands.
Not sure which limits to use? See server monitoring alert thresholds. More on server monitoring.