Create your first cluster

You create clusters from the Aether portal. A cluster comes up with a fully managed control plane and your worker nodes, ready to run workloads in a few minutes.

Before you create

Three things worth knowing before you submit the form:

• The cluster’s public API endpoint costs €4 per month, and it is mandatory. Every cluster gets one. That flat charge is locked the first time the billing sampler observes the address in the billing month, and it is not covered by any grace period — once the sampler has observed the address, deleting the cluster the same day doesn’t remove the charge. See the Billing model.
• Every node has a 15-minute grace period, measured from the moment it first appears in the cluster’s node list. A node removed within that window carries no node charge — whether or not it becomes Ready, and whether it was one of the pool’s first nodes, one added later, or a replacement. Past that, the pool’s billing mode applies: a fixed pool bills per pool (its month is not charged if the pool is removed inside its grace period, and is locked once it isn’t), an autoscaling pool bills per started node-hour, counted from that same first appearance in the node list rather than from the end of the grace period. The time a node spends draining after you ask for its removal is never billed. See the Billing model.
• The portal isn’t the only way in. Everything here is also available over the REST API. A CLI and a Terraform provider are planned and not available yet.

Create a cluster

1. Sign in to the portal and open Clusters.

2. Click Create cluster and fill in:

   • Name — a short, descriptive name for the cluster. Because your cluster’s API endpoint and its Kubernetes objects are derived from it, the name must be at most 40 characters of lowercase letters, digits, and hyphens, starting and ending with a letter or digit.

   • Region — currently eu-fi (Helsinki).

   • Kubernetes version — pick from the supported versions. You can give just a major.minor version to get its latest patch, or pin an exact major.minor.patch.

   • Package — a pre-priced bundle of included capacity, or Custom to be billed purely per resource. The four packages, with the whole-cluster allowance each one includes:

     Package	Price	Included allowance
     Dev	€29 / month	1 node · 2 vCPU · 8 GB · 50 GB disk
     Starter	€79 / month	2 nodes · 8 vCPU · 32 GB · 200 GB disk
     Growth	€179 / month	3 nodes · 24 vCPU · 96 GB · 600 GB disk
     Scale	€279 / month	5 nodes · 40 vCPU · 160 GB · 2500 GB disk

     Anything above the allowance is charged per resource as an overage. The allowance is a whole-cluster budget of vCPU, memory and node disk — the node count in the table is illustrative: node count is neither priced nor capped, so spend the allowance across as many or as few nodes as you like. The choice is fixed for the cluster’s lifetime — there is no way to switch a cluster between a package and custom billing afterwards. See the Billing model.

   • First node pool — a name, the exact per-node size (vCPUs, memory, disk), and how many nodes. This is where your workloads will run.

3. Submit the form. Aether begins provisioning immediately.

What Aether does behind the scenes

When you submit the form, Aether:

• Provisions and starts your cluster’s managed control plane — you never have to operate or maintain it.
• Installs the Cilium CNI, so networking and NetworkPolicy support are ready before your workloads land.
• Brings up your worker nodes from the node pool you configured.

Watch it come up

The cluster detail page streams a live activity / events feed while provisioning runs. You’ll see each step as it happens — control plane up, CNI installed, nodes joining — in real time.

When provisioning finishes, the cluster’s status becomes running. That’s your signal that the control plane is healthy and your nodes have joined.

If create fails

Two errors are worth recognising:

• Name conflict — another cluster in your organization already uses that name. Pick a different one and submit again.
• Region capacity — the region has no free capacity left to allocate. This isn’t something you can retry your way out of; open a support request.

Note

Provisioning typically takes a few minutes. There’s nothing to do while it runs — leave the page open to follow the events, or come back later.

Next steps

• Node pools — size and manage your worker nodes.
• Connect with kubectl — download your kubeconfig and start deploying.