Connect

Hosts

Organise SSH hosts into groups and tags, read the live dashboard and Health badges, edit connection settings, and see how production hosts are treated.

The Hosts view is your address book and your dashboard. Every server you connect to is a host here, with its login, its route and a live view of how it is doing.

The host list

Hosts are grouped by their Group field and sorted by name. Hosts without a group sit under Hosts.

This computer is pinned above them. Click it to open a shell on this computer, or its chevron to pick one. See Local terminal.

FilterShows
AllEvery host
OnlineHosts with a live session
ProductionHosts whose group name contains “production”
RecentHosts you used recently, newest first, read from the audit log

Small badges mark a host that routes through a jump host and a sample host. An amber dot means a host-facts warning you have not seen yet (see Host facts).

The search box filters the list as you type. It also adds hosts: paste an ssh … command or an ssh:// link, or type add 10.0.3.4 as deploy behind edge-bastion, and click Add as a host. See Your first connection.

The host dashboard

Select a host to see its dashboard. Double-click to connect.

The header shows the host’s name and status (Online, Connecting, Error or Offline), its ssh user@host line, its group and tags. The buttons are Edit, SFTP and Connect (Open terminal when a session is already live). When a connect failed, Hop Doctor’s headline sits under it as a one-line hint.

While the host is connected, the dashboard shows:

CardWhat it shows
Metric tilesCPU (with the core count), Memory, Disk (the root filesystem), Load (the 1-minute average, against the core count), Processes and Threads, each with its live reading and a sparkline
Load averageThe 1, 5 and 15-minute load, with a chart of recent samples
ConnectionAddress, user, authentication methods, jump host, OS, uptime, shell, and network traffic down and up

When the host is not connected, the dashboard says what it was at the last read, for example its OS, and when you last connected. Live metrics appear once a session is open.

Connected or not, once the host has been read, the dashboard also holds:

  • The Host dossier card. It opens on an overview: the system in one line, then tiles for containers, listening ports, failed units, the fullest disk and the last read. Below it, the facts sit in sections (System, Services, Network, Storage, Containers and Notes), opening on the one that most needs a look. Long lists show their seven most urgent rows, with Show all for the rest and a filter once there are more than ten. See The Host dossier.
  • The Since you were last here card, with the most notable changes first.
  • The Connections card, from the service map.

While the host is connected, Ask Farabi on a live stat tile asks about that figure in the host’s terminal, and the host’s row in the list carries a badge when Health has raised a signal.

Recent activity lists the last audit entries about this host, and Notes shows what you wrote in the host editor.

How metrics are read

Metrics are sampled over the session you already have open: one read-only command round trip per interval, capped at 10 seconds and 64 KB. The interval is 5 seconds by default. Change it in Settings › Hosts › Monitoring › Metrics polling; 0 stops polling. top is deliberately not used, since it costs a second per sample.

Load average is shown as saturation against the host’s core count, so a load of 3.4 on 16 cores reads as busy rather than critical.

Edit a host

Click Edit on the dashboard, or create one with New. The editor has four tabs.

TabFields
GeneralDisplay name, Hostname or IP, Port, Username, Group, Tags
AuthenticationWhich methods to try, the private key file, saved secrets with Forget
RoutingJump hosts, in order, and a preview of the route
AdvancedKeepalive, SFTP start directory, Run on shell open, Remember facts about this host, Notes

Press Esc to close without saving and ⌘Enter (CtrlEnter) to save from anywhere in the editor.

Advanced settings

  • Keepalive: seconds between keepalive messages; 0 turns them off.
  • SFTP start directory: where the remote file pane opens, for example /var/www.
  • Run on shell open: a command sent once as soon as the interactive shell is ready, for example cd /srv && tail -f app.log.
  • Remember facts about this host: unticked, no read-only checks run on connect and nothing is stored. Unticking it also forgets what was already kept.
  • Notes: anything worth remembering. Notes also appear in the dossier.

Groups and tags

Use Group for where a host lives (Production, Staging, a customer) and Tags for anything else. Type a tag and press Enter; Backspace in an empty tag field removes the last one.

Production hosts

Gatesys SSH treats a host as production when its group or any of its tags contains “prod”, for example Production, prod-eu or db-prod. Production hosts get extra care:

  • Host facts asks once, in the terminal, before its first read.
  • Safe Change requires the safety net for sshd, sudoers, PAM and firewall files, unless you type the host’s name.
  • SFTP writes to those files outside the review sheet are refused.
  • A tunnel described in a sentence connects read-only first where the service supports it. See Port forwarding.
  • A command Farabi suggests that is rated Danger becomes Critical, so Run asks you to type the host’s name. See Command safety.

Delete a host

Deleting a host also removes it from other hosts’ jump chains, forgets its saved password and passphrase, and deletes the facts kept about it.

Sample hosts

Sample hosts appear only when you skip the first-run setup’s Servers step with an empty host list. See Sample hosts. They are ordinary, editable hosts pointing at addresses that go nowhere. Editing one makes it yours. Remove sample hosts at the bottom of the list clears the rest.

Something unclear or wrong? Tell us.