Compute

The Compute group covers everything that runs workloads in the Enterprise cloud. Every page applies to the region in the top-bar region selector.

Service	Purpose
Virtual Instances	Provision and manage virtual machines.
Bare Metal	Dedicated physical servers.
Kubernetes	Managed clusters and their node pools.
Databases	Managed PostgreSQL clusters. In beta.
Containers	Run a public container image and scale it on demand.

Virtual instances

The page shows the total number of instances across all regions, and how many are running and stopped. Select Create instance to open the creation form.

Section	Choices
Instance name	A name for the VM.
Image	A platform image — Ubuntu, Debian, CentOS, Fedora, Rocky Linux, Windows, openSUSE, Fedora CoreOS, FortiGate or MikroTik CHR, each in several versions — or one of your custom images. The form shows the default login user, for example ssh ubuntu@ip-address.
Type	Standard, CPU optimised or Memory optimised, then a size from 1 vCPU / 2 GiB up to 32 vCPU / 128 GiB.
Volumes	A system volume (20 GiB by default) plus any extra volumes, each with a name, size in GiB, type (for example standard or ssd_hiiops) and an optional attachment tag.
Network settings	One or more interfaces, each Public or Private. Private networking is available once you have created a network.
Firewall settings	The security group to apply. Every region starts with a default group.
SSH key	One of your Enterprise SSH keys.

Bare metal

Bare metal servers are dedicated physical machines. The page counts total, running and powered-off servers. Select Create server to open the creation form.

Section	Choices
Server name	A name for the server.
Image	A platform image — Ubuntu, Debian, CentOS, Rocky Linux or Windows — or one of your custom images. The form shows the default login user.
Server type	High frequency or Infrastructure. Each configuration card lists the CPU, memory, disks and RAID layout, and network ports, with the number of servers in stock. Show unavailable reveals configurations that are out of stock.
Network settings	One or more interfaces, each Public or Private.
SSH key	One of your Enterprise SSH keys. Create key adds one without leaving the form.
Additional options	User data runs a cloud-init or shell script on first boot. Add tags stores metadata, and the keys are mirrored onto the machine.

The Summary panel shows the region, image, configuration, disk and network. It checks the request against your quota once you have chosen an image and a configuration.

Note

Physical servers cannot take floating or reserved IP addresses. A public interface uses the address that the provider assigns.

Kubernetes

Managed Kubernetes clusters and their node pools. Select Create cluster to open the creation form.

Cluster

Field	Notes
Name	A name for the cluster.
Kubernetes version	One of several supported minor versions. The newest is selected by default.
SSH key	Installed on every node. Create key adds one without leaving the form.

Node pools

A cluster starts with one pool, pool-1. Select Add pool to add more, for example to separate GPU workloads from general ones.

Field	Notes
Pool name	A name for the pool.
Minimum nodes, Maximum nodes	The range within which the pool can scale.
Type	Virtual instances or Bare metal instances, then a shape.
Volume type, Size	The volume attached to each node. Every node gets its own volume, and each one counts against your volume quota.
Placement policy	Anti-affinity soft (the default), Anti-affinity or Affinity. See placement groups.
Labels, Taints	Optional Kubernetes labels and taints applied to the pool’s nodes.
Autohealing nodes	Replaces a node that stops reporting healthy.
Public IPv4 address	Required while the cluster has no private network to reach the nodes on.

Network and advanced settings

Field	Notes
Private network	A network for the nodes. With None, public addresses, every node takes a public address.
CNI	Calico or Cilium.
Cluster auto-healing	Replaces control-plane nodes that stop reporting healthy.
Cluster autoscaler	Tunes how the cluster adds and removes nodes.
Set internal IP ranges	Use this only when the cluster’s internal ranges must not overlap an existing network.

The Summary panel shows the region, version, number of pools, nodes at start and network type. It checks the request against your quota once every node pool has a shape.

Databases

Managed PostgreSQL clusters. This service is in beta and is offered in fewer regions than the rest of Compute, so the page may open with a different region selected. Select Create cluster to open the creation form.

Cluster and resources

Field	Notes
Name	A name for the cluster.
Compute	From 1 vCPU / 1 GiB up to 8 vCPU / 16 GiB.
PostgreSQL version	PostgreSQL 13, 14 or 15.
Volume size, Volume type	Storage in GiB, on ssd-hiiops, ssd-lowlatency or standard.

Network and access

Field	Notes
Network interface	Public or Private.
Access control list	Addresses or CIDR ranges allowed to connect on TCP port 5432. The default, 0.0.0.0/0, allows every address.

Caution

Replace the default 0.0.0.0/0 entry with your own address ranges before you create a public cluster. Otherwise the database port is open to the whole internet.

DBMS settings

Setting	Notes
High availability	Adds a replica. Without it, the cluster is a single instance.
Connection pooler	Runs PgBouncer in front of the cluster.
postgresql.conf	Tuned by the provider for the chosen CPU and memory.

Users and databases

The form requires at least one user and one database.

• Users need a username in lowercase letters, and any of the role attributes LOGIN, NOLOGIN, INHERIT, CREATEDB, CREATEROLE and BYPASSRLS.
• Databases need a name and an owner, chosen from the users above. Add users first.

The Summary panel shows the region, version, compute, storage and replication mode, and checks the request against your quota once a user and a database are named.

Containers

Run a public container image without managing servers, and scale it with demand. Select Create container to open the creation form.

Section	Field	Notes
Container	Name	Also used to build the container’s public hostname.
Image and port	Image	A public container image. Required.
	Port	The port inside the container that traffic is sent to. The default is 80.
	Startup command	Optional. Overrides the image’s own entrypoint.
Configuration	Size	From 80 mCPU / 128 MiB up to 2260 mCPU / 4096 MiB.
	Minimum replicas, Maximum replicas	The scaling range. A minimum of zero lets the container scale to nothing between requests.
	Cooldown	Seconds to wait between scaling actions.
	Autoscaling triggers	The conditions that add replicas. Without a trigger, the container stays at its minimum.
Settings	Environment variables	Passed to the container when it starts.
	Request timeout	How long a single request may run before it is cut off.

The Summary panel shows the region, image, size, replica range and port, and confirms whether the container fits within your quota in the selected region.

Tip

A minimum of zero replicas is the cheapest option for occasional traffic. However, the first request after an idle period waits while a replica starts. Set the minimum to one for latency-sensitive services.