Live dashboard for core services on the host: state, CPU, RAM, and uptime, with one-click controls (start/stop/restart/reload). Access: root.

What it shows

• Status — Active / Inactive / Failed / Not installed (from systemctl).
• CPU / RAM — point-in-time % for the unit’s MainPID (via ps).
• Uptime — time since the unit entered the “active” state (derived from ActiveEnterTimestamp).

How services are discovered

• Reads services.yaml and builds a list of enabled components.
• Special cases:
  • nginxapache: install=on → adds both apache and nginx.
  • database → adds mariadb, and phpmyadmin if enabled.
  • filetransfer → adds ftp plus helper daemons pure-certd and pure-authd; adds sftp if enabled.
  • mail → adds mail (Exim), dovecot, spamassassin, and roundcube if enabled.
  • Per-user PHP pools are intentionally omitted.

Unit mapping

Friendly names are mapped to systemd units, e.g. apache→httpd, nginx→nginx, mariadb→mariadb, ftp→pure-ftpd, mail→exim, etc.

Controls & refresh

• Per-row actions: Start, Stop, Restart, Reload. Each runs systemctl <cmd> <unit> and then refreshes that row.
• Refresh status (row) — re-polls a single service.
• Refresh all — reloads the full table.
• Badges: green = Active, red = Inactive/Failed, gray = Not installed, amber = other states.

How status is computed

• Installed? systemctl list-unit-files … for the mapped unit.
• State: systemctl is-active.
• Main PID: systemctl show -p MainPID → ps -p <pid> -o %cpu=,%mem=.
• Uptime: systemctl show -p ActiveEnterTimestamp → human-readable elapsed time.

Notes & tips

• “Not installed” can also mean “no unit file present for the mapped name.”
• CPU/RAM reflect the main process only; auxiliary workers are not summed.
• Some stacks run multiple units (e.g., FTP helpers); each is tracked individually.