Skip to content
zefıly

Servers

The server page

Overview, health score, live status, security baseline, automatic agent updates, reboots, reinstalling and what an offline server shows.

Open a server from Servers to manage it. The left menu groups its sections:

Group Sections
General Overview, Activity
Apps & data Sites, Web stack, Databases, Backups
System Performance, Updates, Services, Firewall, Network, Users, Cron, Daemons, Disk
Tools Files, Terminal

A server opens on its Overview; a link to a section (from the home page, say) opens that section, and the page stays on the section you are working in when an action reloads it.

The header shows the server's state (Online, Offline, Waiting for agent), its address, system and health score, with Terminal and Reboot buttons.

Work in progress

Whatever you start on a server — installing a database server, the web stack, PHP versions or extensions, setting up a site or an app, a deploy, an agent update — shows in a strip at the top of the server page while it runs: what it is, whether it is waiting for the server, and once it runs the step it is on ("Refreshing package lists", "Installing 16 package(s): …", "PHP 8.5: 16 packages"), and for how long. When it finishes, the strip shows done or why it did not finish for a few seconds. Jobs waiting for your passkey show there too, with an Approve button. Jobs that never report back (the agent restarted mid-way) are closed after their time limit, so nothing waits on them.

Overview

  • Load, Memory, Disk and Uptime tiles from the agent's last report (the server reports every 5 minutes; uptime keeps counting in between). Refresh asks the server for a fresh report.
  • History: load, memory and disk over the last day to 90 days (5-minute averages; how long depends on your plan) — see Monitoring and alerts.
  • Live status: memory, swap, every disk, failed services, and a notice when a reboot is needed to finish updates.
  • Server: hostname, operating system, kernel, architecture, CPU cores, memory, disk and agent version.
  • Health score (0–100) with its issues: offline, base setup not run, failed services, reboot required, a disk 80% or 90% full, pending security updates, an outdated agent or system, failing security checks, and an expiring agent certificate.
  • Baseline setup: what the automatic setup did; Run baseline setup runs it again.
  • Security baseline: CIS-lite checks of SSH, the firewall, automatic updates, kernel hardening and accounts (Run security checks), and Check for drift to find files Zefily manages that were changed by hand.
  • Diagnostics: Collect diagnostics gathers web server, PHP and agent logs, package activity and out-of-memory kills when something is wrong.
  • Agent updates: agents update themselves to the newest release of their rollout (checked every 10 minutes, at most every 6 hours per server, never while a job runs; a new agent that does not reconnect within 3 minutes rolls back on its own). When a newer agent is available, a card at the top of the overview offers Update agent to do it right away.

SSH password login

Zefily keeps SSH password login on until an SSH key is set for root or a sudo user, so you can never lock yourself out. Add your key to a sudo user on the Users section, then run the baseline setup again: password login is turned off.

Activity

Every job the panel sent to this server — who asked for it, when, its state and its log. Jobs waiting for your passkey can be approved here.

Reboot

Reboot asks for confirmation and your passkey; sites are unavailable for about a minute and the server reconnects on its own. If packages are being installed at that moment, the server finishes them first and reboots right after (interrupting a package installation would break it). Once the server is back, the message under Reboot goes away; if the reboot could not run, it says why. Servers at a connected cloud provider also have Power-cycle at the provider for a server that does not answer.

Reinstall a cloud server

Servers created through a connected Hetzner, Linode or Vultr account can be reinstalled in place from the provider card on the overview: Reinstall, choose the operating system and type the server's name to confirm.

  • Everything on the server's disk is erased: sites, databases, files and settings.
  • The server keeps its address, plan and additional IP addresses.
  • In the panel it starts fresh like a new server — the old system's sites, databases and jobs are removed, its name, group, tags and chosen web stack stay — and it enrolls and sets itself up on its own within a few minutes.
  • Backups in your own storage stay: restore your sites from them afterwards.

Reinstalling, deleting a server and removing a provider account ask you to confirm it is you (passkey or password) if you have not in the last five minutes; you then repeat the action on the page you return to.

DigitalOcean and AWS cannot reinstall with a new setup: delete the server and create a new one there.

When a server is offline

If the agent has not reported for a while, the server page says This server is offline — last seen …. Live figures, updates, services and tools are paused (their sections are greyed out); the server's record, activity and sites stay visible, and changes run once the agent reconnects. If the machine is running, check the agent on it:

systemctl status zefily-agent

Tags, groups and bulk actions

On the Servers list, tag servers, move them into groups and select several of them for bulk actions: ping, refresh status, check updates or install security updates.

Updated October 20, 2026 View as Markdown llms.txt

Still stuck?

Open a ticket from your panel or write to us; we answer within a day.

Contact →