Using CERMM

Devices

Every machine with an agent, problems first. Click one for its page.

Needs attention

Above the list, only what needs someone: servers down and critical alerts. Each server is its own row; workstations' problems are grouped per rule ("3 workstations: disk over 95%"). CERMM never headlines how many machines are online or offline: at a few hundred machines that number tells you nothing.

The list

Each row has:

• a status light, and an icon for workstation, server, laptop, VM or container;
• the name, its IP address and who's signed in;
• the system's logo with its OS, release and build or kernel;
• make and model, with threads and memory;
• CPU, memory and disk meters, coloured at the alert lines;
• uptime and when it was last seen.

From 1800 pixels wide it adds Updates and Agent columns.

Pills along the top: All, Servers, Workstations, Containers, Needs attention, Any alert, Offline, and your groups. The filter box matches names, addresses and more; #tag matches a tag. Problems sort first, and the list shows 200 rows at a time.

Right-click a row for its menu: Open, Copy IP address, Terminal, Run script…, Install software…, Capture as playbook… (Linux), New ticket, Tags…, Add as asset (or Open asset), Print label…, Mark as server or workstation, and Remove from CERMM.

Workstation or server

CERMM works it out: a laptop is a workstation; Windows Server, a server chassis, or Linux without a desktop is a server. Set it by hand with right-click → Mark as server (or workstation); Work out server or workstation again goes back to CERMM’s guess. Containers (LXC, Docker, Podman, from what the agent reports about virtualization) are under Containers, but still count as servers for alerts and schedules.

The IP address

The address shown is the machine's own primary address: the interface that holds the default gateway. Not the address its reports arrive from, which is a router's or a host's when the machine is behind NAT.

A device's page

The title has the name, the OS logo, and three buttons: Connect, Run script and Install software (see Run). Under it, the nameplate: the facts as label over value, its tags, and Host if it's a guest on a hypervisor CERMM knows.

Under that, tabs. A tab shows up only when the machine has something for it, and the page opens on Overview.

• Overview: gauges for CPU, memory and disk; open alerts and their history, with Alert rules for this machine's own thresholds (Alerts); 24 hours of CPU and memory in ten-minute steps (a stretch with no reports is a gap, never 0%); disks; hardware and system.
• Tickets for this machine (Tickets).
• Updates: pending updates, Check now, Install, and Reboot when one is pending (Patches).
• Software and Network (gateway, DNS, interfaces).
• Virtual machines on a Proxmox or Hyper-V host (Proxmox and Hyper-V), Containers on a Docker or Podman machine (Containers), Backups (Backups).
• Activity: the last jobs; click one for its output.
• Asset, Documentation and Passwords linked to it.
• Agent: version, connection, updates, and Remove device.

Software

Every installed package, from dpkg, rpm, pacman, flatpak and snap on Linux, the Uninstall keys on Windows and system_profiler on macOS. A filter, and 25, 50, 100 or 250 per page (remembered). Click a package for what it is: the one-line summary from the package manager, versions, KB numbers on Windows, a link to its page, and Uninstall.

Offline

A machine that hasn't reported for 90 seconds is offline. That's only an alert for servers (after a few minutes; Settings → Alerts). Workstations go home at night.