Skip to content

Datacenter options

Datacenter → Options holds the defaults that apply across the whole system. Select a row and click Edit to change it.

Most can be left alone. Two are worth changing on the first day, and one can lock everyone out — those are called out below.

Setting Why
Email from address Ships as a placeholder that most mail systems reject. Until it is a real, accepted address, notifications — including backup failures — do not arrive. See Notifications and monitoring.
Keyboard Layout Decides how console keystrokes are interpreted. If it does not match the keyboards actually in use, typing in a console produces the wrong characters — which is baffling when you are entering a password you cannot see.

Datacenter options on a new cluster

Setting What it controls
Keyboard Layout Keyboard layout used by guest consoles
HTTP proxy Outbound proxy for update and licence traffic
Console Viewer Which console technology the Console button opens
Email from address The sender address notifications come from
MAC address prefix The prefix used when generating machine network adapter addresses
Migration Settings Which network migrations use, and whether they are encrypted
Replication Settings Global limits for replication traffic
HA Settings Cluster-wide high availability behaviour, including what happens to protected machines when a node is shut down deliberately
Cluster Resource Scheduling How the cluster chooses a node when placing or rebalancing machines
U2F Settings / WebAuthn Settings Parameters for hardware security keys
Bandwidth Limits Default caps for cloning, migration, restore and backup traffic
Maximal Workers/bulk-action How many machines a bulk action processes at once
Next Free VMID Range The range the “next free ID” suggestion draws from
Tag Style Override Colour and shape of tag labels
User Tag Access Who may apply which tags
Registered Tags The approved list of tags, when tag access is restricted
Consent Text Text shown and accepted at sign-in
Location Geographic label; nodes inherit it unless they override it

By default, migration uses the management network. On a busy or slow management network, moving a large machine can interfere with the traffic the cluster uses to check its nodes are alive. If you have a separate, faster network, point migration at it here.

Sets ceilings for clone, migration, restore and backup traffic across the cluster. Worth setting where those operations share a link with production traffic — a restore that saturates the network is its own kind of outage.

Includes what should happen to protected machines when a node is deliberately shut down — whether they migrate away first, or are simply stopped and restarted elsewhere. Get this right before your first planned maintenance.

Configure this before asking people to enrol hardware security keys. Enrolments are tied to the settings in place at the time, so changing them afterwards can invalidate keys that were already registered.

By default any user who can edit a machine can apply any tag. If you use tags for anything that matters — environment, owner, billing — restrict access here and define the approved list in Registered Tags, so the vocabulary stays consistent.

Anything entered here is shown at sign-in and must be accepted before anyone can proceed.

Several settings have a per-node counterpart under node → System → Options — most visibly Location, which reads from Datacenter options on a node until you override it. See the node menu reference.