Just Dashboard Processes and services
Inspect live processes, PM2 apps, systemd units, and scheduled work.
Processes sits under Monitoring in the sidebar and has four pages: Live, PM2, Services, and Scheduled. It distinguishes host processes by their cgroup or supervisor rather than guessing ownership from an executable name, so a service, container process, shell job, and unmanaged process can point to different remedies.
Find a busy process
Open Monitoring → Processes → Live. The page opens on who is using the machine: the five heaviest workloads by CPU and by memory, where a workload is a supervisor's processes or the copies of one program started by hand. Select one to narrow the table to it. The table's head counts running, sleeping, blocked, and zombie processes; search or filter by user, state, and owner, then sort. The automatic focus chooses CPU, memory, or disk I/O from current host pressure and explains why; you can override it. Open a row for its live CPU, memory, disk I/O, and open files against the limit, its command and working directory, what it listens on, and its parent chain and children. Process environment variables are not exposed in the detail sheet.
The row menu offers Terminate, Kill, Pause/Resume, Hang up, Interrupt, and Send SIGUSR1/Send SIGUSR2 when your role permits them. Signals are confirmed and checked against the process start time, so a PID reused after the page loaded cannot target a different process. PID 1 and the dashboard's own process are refused, and no signal is offered for a kernel thread or a zombie; a zombie's menu opens its parent instead. Changing priority requires an administrator; do it only when you understand how it affects other work on the host.
Manage supervised work
PM2 discovers application daemons for host accounts. Its first line shows the account, the Node version, the boot hook, when the startup list was last saved, and whether the applications come back after a reboot. Inspect each app, use Start, Reload, Restart, or Stop, open its logs, or use Scale… to change a cluster's size. Reset restart counter, Flush logs, and Delete from PM2 are in the row menu. Start application accepts a script or ecosystem file and requires an administrator because it executes host code. A running PM2 app is not necessarily ready after reboot: an app started since the last save turns the line amber and offers Save list. Saving the list does not install a missing boot hook.
Services lists systemd units, failed first, with their state, startup setting, and CPU and memory. Its detail sheet shows live readings, the command, how the unit runs, and a Journal tab whose Runs view groups a week of the unit's starts and stops. Start, Stop, Restart, Reload configuration where the unit supports it, Enable on boot/Disable on boot, and Clear failed state are available; after changing a unit file, use Reload unit files. A static unit has no enable switch. sshd's unit journal is shown only to administrators.
Scheduled draws the next 24 hours of cron jobs and systemd timers as lanes, with a countdown to the next run. Choose an account to see its crontab, with schedules written in plain words. Next runs use the host's time zone, or the job's CRON_TZ, when it is known and are shown in your browser's local time; otherwise they are estimated in the browser's zone. Add job, Edit, Enable/Disable, or Remove a job. System cron files are shown read-only. Timers offer Run now, Start timer/Stop timer, and Enable on boot/Disable on boot. A cron job has no Run now action: its command is a shell string and is not executed from this page. The Cron log panel shows what the cron daemon recorded; use Logs for anything else scheduled or supervised work wrote.