Quickstart

This document provides a step-by-step guide to deploying and optimizing a cluster with CloudPilot AI. It covers cluster creation, connection to the CloudPilot AI console, installation of optimization components, and cluster rebalancing.

Deployment Options

CloudPilot AI offers two deployment methods:

Console-Based Deployment (This Guide)

Follow the step-by-step instructions below to deploy CloudPilot AI through the web console. This method is ideal for:

• Quick setup and testing
• Manual cluster management
• Learning CloudPilot AI features

Infrastructure as Code with Terraform

For production environments and infrastructure automation, use the CloudPilot AI Terraform Provider. This method provides:

• Automated deployment and configuration
• Version-controlled infrastructure
• Repeatable and consistent deployments
• Integration with existing Terraform workflows

Choose the deployment method that best fits your workflow. The following steps guide you through the console-based deployment.

Prerequisites

• A supported Kubernetes cluster: Amazon EKS, Alibaba Cloud ACK, or Google Kubernetes Engine (GKE).
• Administrative access to the cluster and to the corresponding cloud provider account.
• kubectl installed and configured to talk to the cluster.
• For EKS: awscli and eksctl.
• For ACK: the aliyun CLI.
• For GKE: gcloud, helm, jq, and curl, plus a cluster with Workload Identity enabled before Phase 2.
• Access to the CloudPilot AI console .

Step 1: Create or Select a Supported Cluster

CloudPilot AI currently supports EKS (AWS), ACK (Alibaba Cloud), and GKE (Google Cloud) clusters. If you already have a compatible cluster, skip this step.

• To quickly create an EKS cluster, refer to the Demonstration Cluster for CloudPilot AI .
• To quickly create an ACK cluster, refer to the ACK examples used throughout the Alibaba Cloud provider docs.
• For GKE, use your standard GKE provisioning workflow first.

Step 2: Connect Your Cluster to CloudPilot AI

After logging into the CloudPilot AI console , perform the following:

1. Click Add Cluster.
2. Copy and execute the provided shell command in your terminal.

After executing the script:

1. Click I ran the script.
2. Click the selected cluster, navigate to the Cluster and Cost Monitoring tab, and view the potential savings.

GKE note: the Add Cluster script uses CLOUD_PROVIDER=gcp and relies on GKE metadata during Phase 1.

Step 3: Install CloudPilot AI Optimization Components

Once your cluster is connected, proceed to install the optimization components:

1. Run the provided shell script in your environment.
2. After successful execution, click I ran the script.

You will be redirected to the Rebalance page.

GKE note: the GKE Phase 2 flow requires the exact CLUSTER_LOCATION and a cluster with Workload Identity enabled. Use a zone such as us-central1-a for zonal clusters and a region such as us-central1 for regional clusters.

Step 4: Rebalance Your Cluster

To optimize your cluster, initiate the rebalance process. This consists of three steps:

1. Launch new replacement nodes.
2. Drain workloads from old nodes to the new ones.
3. Terminate the old nodes safely after draining.

This ensures zero-downtime workload migration while optimizing resource usage.

Once completed, your cluster will be under CloudPilot AI management. It will automatically scale out during workload surges and scale in during low usage, always selecting the most suitable node types for your workloads.