> For the complete documentation index, see [llms.txt](https://docs.duplocloud.com/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.duplocloud.com/docs/marketplace-agents/aws-extension/quick-start/provision-new/step-2-cluster.md).

# Step 2: Create a Cluster

Create a Cluster Baseline — an EKS cluster built on top of your Network Baseline.

A Cluster Baseline provisions an EKS cluster inside the Network you created in the previous step. The cloud account, region, VPC and subnets are inherited automatically from the Network Baseline.

## What gets created

* An EKS cluster with your chosen Kubernetes version
* The EKS service IAM role, a cluster security group and an all-host security group
* A default system node group in the network's private subnets (if **Add Default System Node Group** is enabled)
* An OIDC identity provider for IAM Roles for Service Accounts (IRSA)
* A Kubernetes Provider and Scope automatically connected to the new cluster

## Walkthrough

### Step 1 — Navigate to Cluster

In the left sidebar, open **Infrastructure Agent** and select **Cluster**.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-264e0774ae491bbea42891044b933dc581cc2444%2Fcluster-step-01.png?alt=media)

### Step 2 — Cluster List

The **Clusters** page lists every cluster in the workspace, with summary cards for total, ready, in-progress and needs-attention clusters. Each row shows the cluster's cloud, network, type, Kubernetes version, status and mode (**Create** or **Import**). Use the cloud filters to narrow the list. Click **Create cluster** to begin.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-aff48acd92dd5806867d9b8125f73e83fca2d3d4%2Fcluster-step-02.png?alt=media)

### Step 3 — Cluster Details

The **Create Cluster Baseline** wizard opens. Fill in the first page:

* **Mode** — **Create new** to provision a new EKS cluster, or **Import existing** to register a cluster that already exists
* **Cloud** — select **AWS**
* **Name** — a unique identifier for the cluster (e.g. `prod-eks-cluster`). Letters, numbers and hyphens.
* **Description** — optional
* **Metadata** — optional key/value pairs to attach to the cluster

Click **Next**.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-efcf09a75e64dc9faf1103b009e06d5e1778050b%2Fcluster-step-03.png?alt=media)

### Step 4 — Cluster Spec

Define the cluster configuration:

* **Network Source** — select **Choose Network** to inherit the cloud account, region, VPC and subnets from a provisioned network. **Choose VPC** lets you enter an account, VPC ID, region and subnet IDs directly instead.
* **Network** — the network you created in Step 1 (e.g. `prod-network-1`)
* **Cluster Type** — `Standard`, or `Auto` for EKS Auto Mode
* **Version** — the Kubernetes version to deploy. Defaults to the latest stable release once a network is selected.
* **API Server Visibility** — `Public` exposes the Kubernetes API endpoint publicly, `Private` restricts it to the VPC, and `Public and Private` enables both
* **Control Plane Logging** — optional; select which EKS control plane log types to enable (API Server, Audit, Authenticator, Controller Manager, Scheduler)
* **Cluster IP CIDR** — optional; the CIDR block for Kubernetes service IP addresses. Leave blank to use the EKS default.

If the selected network has no NAT gateway, the wizard warns you: EKS nodes need outbound internet access to join the cluster, so pick a NAT-enabled network before provisioning.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-2f88b34f222e74e90ed5ccc56b0efd2678250747%2Fcluster-step-04.png?alt=media)

### Step 5 — System Node Group and Domain Filter

Scroll down to finish the spec:

* **Add Default System Node Group** — on by default (Standard clusters with **Choose Network** only). Provisions a managed node group named `<cluster>-system-nodes` in the network's private subnets. The cluster needs it to schedule EKS add-ons such as CoreDNS and kube-proxy.
  * **Instance Type** — an x86\_64 instance type (e.g. `t3.medium`). ARM and GPU families aren't offered.
  * **Min Size** / **Max Size** — the node group's scaling bounds (default `2` and `3`)
* **Domain Name Filter** — optional; select one or more Route 53 hosted zones from the cloud account. When set, External DNS is enabled on the cluster automatically once it's Ready, scoped to these zones.

Click **Create & Provision**.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-93133d19c6df2f64b3a723a8e04185bfe97630dd%2Fcluster-step-05.png?alt=media)

### Step 6 — Cluster Detail Page

DuploCloud creates the cluster and opens its detail page. The header shows the account and, once provisioned, the cluster ARN. Use the **Spec** / **Result** toggle to switch between the configuration you submitted and the provisioned cluster details. The **Provisioning** panel on the right tracks the cluster's status, and **Configuration** shows the cluster type and network.

Provisioning runs as an agent ticket: click **Track Status** (or **Ask the Agent** in the header) to open it.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-430d32503e0c2cee4bdfd0feba9eedcc3fa23405%2Fcluster-step-06.png?alt=media)

### Step 7 — Agent Ticket Opens

The agent uses the `duplo-aws-infra` skill and begins by reading the `aws-cluster-baseline.md` skill file to load its instructions. It then checks its environment, sets up AWS credentials for the target account and loads the cluster spec.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-46f4272fcdd8709a1ed5cb94b09bd4a9730e43e2%2Fcluster-step-07.png?alt=media)

### Step 8 — Phase 1: Parse & Validate

The agent reads the cluster spec and validates each field before touching AWS:

* Name, VPC, region and subnets — inherited from the linked Network Baseline
* Cluster type, Kubernetes version and API server visibility
* Cluster IP CIDR and control plane logging, when set
* At least two subnets are available for the cluster
* The system node group's instance type and min/max size, when enabled

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-141decad4ee89eeede544248069c18aba5db2fe0%2Fcluster-step-08.png?alt=media)

### Step 9 — Phase 2: Compute

**Phase 2 — Compute** derives the resource names and counts without making any AWS calls:

* The CloudFormation stack name and the EKS service IAM role name, both derived from the cluster name
* Whether an Auto Mode node role is needed
* The total number of resources to create

Then **Phase 3 — Provision** begins. Step 4.1 creates the Auto Mode node role (skipped for Standard clusters), and Step 4.2 starts deploying the CloudFormation stack.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-71c9aa792e08b21aa31c2a56ca7b059ec8ee98b4%2Fcluster-step-09.png?alt=media)

### Step 10 — Deploy the CloudFormation Stack

**Step 4.2 — Deploy CloudFormation stack**:

* Copies the CloudFormation template and builds its parameters
* Checks whether the stack already exists to prevent duplicate provisioning
* Creates the stack — the EKS cluster, its IAM roles, security groups and, when enabled, the system node group — and follows the stack events until it completes (around 15 minutes)

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-3f9372e79979b0b70c9198bb44cd9f6a95236038%2Fcluster-step-10.png?alt=media)

### Step 11 — Stack Outputs and OIDC Provider

Once the stack completes:

* **Step 4.2.5** — Reads the stack outputs: cluster ARN, API endpoint, certificate authority data, OIDC issuer and security group IDs, and posts a partial result to DuploCloud
* **Step 4.3** — Creates the OIDC identity provider, registering the cluster's OIDC issuer as a trusted identity provider in IAM so workloads can use IAM Roles for Service Accounts

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-7b3fd83afbf97d1c05f8a571cd75658b3f92c4a4%2Fcluster-step-11.png?alt=media)

### Step 12 — Provisioning Complete

**Phase 5 — Output** finishes the run:

* **Step 5.1** — Writes `output.json`, used later for de-provisioning
* **Step 5.2** — Posts the final results and sets the cluster status to **Complete**. DuploCloud then connects a Kubernetes Provider and Scope to the new cluster.
* **Step 5.3** — Prints the completion summary: cluster name, type and version, cluster ARN, API endpoint, OIDC issuer URL and provider ARN, and the cluster and all-host security groups

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-bd724236d84e4dc4ea3da22a34282ab1ca6340aa%2Fcluster-step-12.png?alt=media)

### Step 13 — Cluster Result

Back on the cluster detail page, the status shows **Ready**. Select **Result** to see the provisioned cluster's identifiers:

* **Cluster ARN** — the Amazon Resource Name for the EKS cluster
* **API Endpoint** — the Kubernetes API server URL
* **OIDC Issuer URL** — used to configure IAM Roles for Service Accounts (IRSA)
* **Certificate Authority** — the base64-encoded CA data for `kubectl` configuration

The **Actions** menu offers **Kubectl**, **Download Kubeconfig**, **Deprovision Cluster** and **Deprovision Attributes only**. The cluster's left-hand navigation gives access to **Attributes** and to its Kubernetes resources: **Workloads** (Pods, Deployments, StatefulSets, DaemonSets, Jobs, CronJobs), **Configs**, **Network**, **Volumes** and **Observe**.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-23aa3d5b08b7f254ebfda18b4957f565a41dbb7d%2Fcluster-step-13.png?alt=media)

### Step 14 — Cluster List: Ready

The Cluster list now shows the cluster with **Status: Ready**. The three-dot menu on a row offers **View** and **Deprovision**. Deprovisioning an AWS cluster tears down everything the platform created for it; **Delete** is only offered for Kubernetes-only clusters, which register an existing cluster without provisioning anything.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-3e5b54f7e7e5270622c9d6e162875bc6653b550c%2Fcluster-step-14.png?alt=media)

### Step 15 — Cluster Attributes

Click **Attributes** in the cluster's left-hand navigation to manage optional Kubernetes components and EKS add-ons. On a new cluster this opens the **Configure Cluster Attributes** form; once attributes have been applied, it shows the attributes object (named `<cluster>-attributes`) with its own provisioning status. The **Components**, **AddOns** and **External DNS** tabs list what's installed.

The available components are:

* Cluster Autoscaler
* Secret CSI Driver
* ALB Load Balancer Controller
* EFS Volumes
* Metrics Server
* Kube State Metrics
* Flux CD
* External DNS

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-c6f14230aab8608fda409aeb1bcf98c001decccf%2Fcluster-step-15.png?alt=media)

### Step 16 — Selecting Components

Check the components you want enabled on the cluster. On a new cluster, select them in the **Configure Cluster Attributes** form and click **Configure**; to change existing attributes, choose **Actions** > **Edit**, then **Save**. Components can be enabled incrementally — you don't need to select all at once.

The **EKS Addons** section installs managed AWS add-ons: click **+ Add Addon** and select one.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-e7c44eaaca9af1b6c3ac7f488a4973fbbb9de396%2Fcluster-step-16.png?alt=media)

### Step 17 — External DNS Domain Filters

When **External DNS** is enabled, a **Domain Filters** field appears and needs at least one domain. Enter the Route 53 hosted zone domain (e.g. `example.com`) and click **+ Add Domain** for additional domains. External DNS only manages records within the specified domains.

Saving the attributes starts an agent ticket that installs the selected components into the cluster. Click **Track Status** to follow it.

![](https://2471407984-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F68cb0s9ce5UIUKWPuYs8%2Fuploads%2Fgit-blob-f5382a69dfa19258d4ca616d89a3b49697475da0%2Fcluster-step-17.png?alt=media)

## Next step

Once the cluster's status shows **Ready**, proceed to [Step 3: Create a Plan](https://docs.duplocloud.com/docs/marketplace-agents/aws-extension/quick-start/provision-new/step-3-plan.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.duplocloud.com/docs/marketplace-agents/aws-extension/quick-start/provision-new/step-2-cluster.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
