Door console¶

Manage DKS clusters in the Door console at https://door.cloud/dks/clusters. Sign in with your Door account. The UI uses your session; you do not paste an API token.

Who can do this: member can open lists and detail (read). Create, scale, exposure, Elastic IPs, firewall, and delete require super_admin. A project admin counts as org member on DKS.

The console refreshes cluster status for you. Cluster detail URLs look like https://door.cloud/dks/clusters/<id>/…. Tabs: Overview, Nodes, Networking, Access & kubeconfig, Settings.

Clusters list¶

Clusters list in the demo organization

URL: https://door.cloud/dks/clusters

Shows

Cluster name (link to overview)
Status chip from the public phase (Requested, Provisioning, ControlPlaneReady, NodesReady often grouped as provisioning; Provisioned often labelled Ready; Failed; Deleting)
Kubernetes version, zone
Node count from pool desired size (not always live Ready workers)
Capacity totals (total_vcpu, total_memory_gib) when present
Age from created_at

Filter pills (All / Ready / Provisioning / Failed / Not Ready) are client-side on that mapped status. Confirmed from the console contract.

Actions

Open a cluster (the name is a button)
+ New Cluster
Refresh the list
Per row: Analyze with Maya, and a … actions menu

The organization name is the heading at the top of the sidebar. Switch organization from the account menu at the bottom of the sidebar.

Maps to GET /v1/public/clusters in Cluster lifecycle.

Create cluster¶

Create a cluster, general information

Node pools step: pool name, replicas, machine type

URL: https://door.cloud/dks/clusters/new

The wizard has three steps: 1. General Information (name, zone, Kubernetes version), 2. Node Pools (pool name, replicas 1–200, machine type cards with vCPU, memory, disk, and an estimated price), 3. Review & Create. Cancel leaves without creating anything.

Shows a form seeded from GET /v1/catalog?zone=abidjan:

Cluster name (DNS label)
Zone (today abidjan)
Kubernetes version (catalog; Abidjan offers v1.32.4)
Node pools: name, machine tier (for example dks.c5.xlarge), count, optional autoscaling
Kubernetes API exposure: Private vs Public (see API server exposure)
Where applications can be reached: toggle Private (VPN) vs Internet (maps to network_mode intranet / internet). Confirmed card title: Where can your applications be reached? Set once; cannot be changed after creation. Required for public Elastic IPs (internet only).

You do not pick a control plane. Door applies the one assigned to your organization (the zone default unless Door set another).

Actions: submit create. On success the console goes to the Provisioning page. Maps to POST /v1/public/clusters (201). Validation errors stay on the form (400 / 409 / 422 / 402 prepaid credit).

Provisioning¶

URL: https://door.cloud/dks/clusters/<id>/provisioning

Shows the five-step create narrative from the cluster progress block (render the API text as-is):

Step	Label you should see
1	Request received
2	Validating configuration
3	Creating infrastructure
4	Control plane ready, then Worker nodes ready
5	Cluster ready

Header badge uses public phase, not a raw internal step. Progress percent comes from progress.percent when set. On failure, the page shows failure_message.

Actions: wait; open Cluster details when the cluster is Provisioned (some builds also enable kubeconfig at Control plane ready). Maps to GET /v1/public/clusters/{id} and GET …/state.

Cluster detail shell¶

URL: https://door.cloud/dks/clusters/<id>/…

Header: name, status, id, version, zone, age. Tabs:

Tab	Path suffix	Purpose
Overview	(index — no suffix)	Identity and capacity
Nodes	`/nodes`	Pools and workers
Networking	`/networking`	Application Elastic IPs
Access & kubeconfig	`/access-kubeconfig`	API URL and kubeconfig
Settings	`/settings`	Tags, description, delete

While the cluster is still provisioning, the console may redirect to the Provisioning page.

Overview¶

Cluster overview

Shows: version, zone, network (VPN / Direct Connect when the cluster is intranet, otherwise Internet), vCPU, memory, storage, created time, and age. The header badge is Ready when the API phase is Provisioned.

The public API does not expose a control-plane "profile" field. If the console still shows a Profile row, treat it as informational.

Actions: none beyond navigation. Data from GET /v1/public/clusters/{id}.

Nodes¶

Nodes tab: one pool and its workers

Shows: one card per node pool (name, replica count, machine type, vCPU, memory, disk) and a worker table (node name, status, disk, CPU, memory, Kubernetes version, age). + Add node pool adds a pool. The row action deletes that node.

Actions (super_admin, cluster Provisioned):

Add pool → Node pools (POST …/node-pools)
Scale / edit autoscaling → PATCH …/node-pools/{name} (size cannot change)
Delete pool (not the last pool) → DELETE …/node-pools/{name}
Delete a node → DELETE …/nodes/{node_name} (Door replaces capacity according to the pool)

Maps to public node-pool and node routes. List nodes is 200 with a snapshot; 409 until Provisioned.

Networking¶

Networking tab: application public IPs

Confirmed in the Door console. One card:

Title Application public IPs — Optional internet-facing addresses for your applications.
Table columns: Node (the Elastic IP name), Status, Public IP (the address), Actions.
Button Add public IP (shown when you can allocate).
Row trash control (aria Delete \<name>) when release is allowed (Ready or Failed).

Empty copy: No public IPs yet. Add one to expose a stable address for ingress.

Maps to Elastic IPs. This tab does not show the Kubernetes API URL (that is Access) and does not include a firewall editor — use the firewall API.

Access & kubeconfig¶

Access and kubeconfig: door CLI, exposure, static kubeconfig

Shows:

Recommended: door CLI — copy-paste commands (doorctl config, doorctl login, doorctl kubectl config --cluster <id>, kubectl get nodes) and Install the CLI. Badge Auto-rotating.
Cluster Access — Private / Public toggle. Private shows API endpoint (VPN / Direct Connect). Public shows the public hostname.
Static kubeconfig (for CI) — Download kubeconfig, unless your organization has disabled it (Managed by your organization policy).

Actions:

Toggle exposure → PATCH …/exposure (202). Wait until the URL matches, then download kubeconfig again.
Download kubeconfig → GET …/kubeconfig (202 then 200). See Kubeconfig and kubectl.

Settings¶

Settings: description, tags, and the danger zone