### Initialize Terraform

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/gke/gke_live_migration/README.MD

Initialize the Terraform CLI in the example root folder. Ensure Terraform is installed and configured.

```bash
tofu init
```

--------------------------------

### Start Minikube Cluster

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/anywhere/README.md

Starts a Minikube cluster with 3 nodes. Ensure Docker Desktop is running.

```sh
minikube start --nodes=3
```

--------------------------------

### Copy Example Terraform Variables

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/gke/gke_clusters_with_workspaces/README.md

Before running the example, rename the provided variable files to their active counterparts.

```bash
cp tf.vars.example tf.vars
cp tf_clusterA.vars.example tf_clusterA.vars
cp tf_clusterB.vars.example tf_clusterB.vars
```

--------------------------------

### Initialize Terraform

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/aks/aks_cluster/README.MD

Run this command in the example root folder to initialize Terraform and download necessary providers.

```bash
terraform init
```

--------------------------------

### Review generated configuration

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/workload_scaling_policy.md

Example of the configuration structure generated by the import process.

```hcl
cpu {
  look_back_period_seconds = 0
}
```

--------------------------------

### Example Usage of castai_workload_scaling_policy_order

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/workload_scaling_policy_order.md

This example demonstrates how to configure the order of workload scaling policies for a cluster. It fetches existing policy IDs, defines a custom order for specific policies, and then concatenates them with the remaining policies sorted alphabetically.

```terraform
data "castai_workload_scaling_policy_order" "cluster" {
  cluster_id = castai_gke_cluster.cluster.id
}

locals {
  ordered_policy_ids = [
    "be61e44b-0f7c-44da-b60e-0594d6fd3634",
    "f60b79f0-d21e-4eda-a59f-f5daa846d289",
  ]
  unordered_policy_ids = tolist(setsubtract(data.castai_workload_scaling_policy_order.cluster.policy_ids, local.ordered_policy_ids))
}

resource "castai_workload_scaling_policy_order" "custom" {
  cluster_id = castai_gke_cluster.cluster.id
  policy_ids = concat(local.ordered_policy_ids, sort(local.unordered_policy_ids))
}
```

--------------------------------

### Deploy Private Hosted Model

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/ai_optimizer_hosted_model.md

This example shows how to deploy a private hosted model. No vLLM configuration is needed for private models.

```terraform
resource "castai_ai_optimizer_hosted_model" "private_example" {
  cluster_id     = castai_eks_cluster.example.id
  model_specs_id = castai_ai_optimizer_model_specs.private_example.id
  service        = "custom-model"
  port           = 8080
}
```

--------------------------------

### Build and Clone CastAI Terraform Provider

Source: https://github.com/castai/terraform-provider-castai/blob/master/README.md

Clone the repository and build the provider locally. Ensure Go is installed and meets the requirements.

```bash
git clone https://github.com/CastAI/terraform-provider-castai.git
cd terraform-provider-castai
make build
```

--------------------------------

### Import castai_node_template resource example

Source: https://github.com/castai/terraform-provider-castai/blob/master/README.md

This command demonstrates how to import an existing `default-by-castai` node template into Terraform state. This is an optional step if following the migration guide.

```sh
terraform import castai_node_template.default_by_castai 105e6fa3-20b1-424e-v589-9a64d1eeabea/default-by-castai
```

--------------------------------

### EKS Cluster Authentication Configuration Example

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_existing/README.MD

This snippet shows an example entry for the aws-auth ConfigMap, which might be necessary if your EKS cluster uses CONFIGMAP authentication mode. It ensures the instance profile used by CAST AI is present.

```yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: aws-auth
  namespace: kube-system
data:
  mapRoles:
  - rolearn: arn:aws:iam::ACCOUNT_ID:role/CAST_AI_NODE_INSTANCE_ROLE
    username: castai-node-role
    groups:
      - system:bootstrappers
      - system:nodes
```

--------------------------------

### Example Reservations CSV Structure

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/reservations.md

Defines the expected structure for the reservations CSV file, including headers and example data for AWS reservations.

```csv
name,provider,region,instance_type,price,count,start_date,end_date,zone_id,zone_name
reservation1,aws,us-east-1,c5n.large,,1,2020-01-01T00:00:00Z,2050-01-01T00:00:00Z,,
reservation2,aws,us-east-1,c5n.large,,2,2020-01-01T00:00:00Z,2050-01-01T00:00:01Z,,
```

--------------------------------

### castai_gke_cluster Resource Example

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/gke_cluster.md

Example usage of the castai_gke_cluster resource to connect a GKE cluster to CAST AI.

```APIDOC
## castai_gke_cluster Resource

### Description

Connects an existing GKE cluster to CAST AI.

### Schema

#### Required

- `location` (String) GCP cluster zone in case of zonal or region in case of regional cluster
- `name` (String) GKE cluster name
- `project_id` (String) GCP project id

#### Optional

- `credentials_json` (String, Sensitive) GCP credentials.json from ServiceAccount with credentials for CAST AI
- `delete_nodes_on_disconnect` (Boolean) Should CAST AI remove nodes managed by CAST.AI on disconnect
- `timeouts` (Block, Optional) (see [below for nested schema](#nestedblock--timeouts))

#### Read-Only

- `cluster_token` (String, Sensitive) CAST.AI agent cluster token
- `credentials_id` (String) CAST AI credentials id for cluster
- `id` (String) The ID of this resource.
- `organization_id` (String) CAST AI organization ID

<a id="nestedblock--timeouts"></a>
### Nested Schema for `timeouts`

Optional:

- `create` (String)
- `delete` (String)
- `update` (String)

### Example Usage

```terraform
resource "castai_gke_cluster" "this" {
  project_id                 = var.project_id
  location                   = module.gke.location
  name                       = var.gke_cluster_name
  delete_nodes_on_disconnect = var.delete_nodes_on_disconnect
  credentials_json           = var.gke_credentials
}
```
```

--------------------------------

### Example Usage of castai_gke_user_policies

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/data-sources/gke_user_policies.md

Use this data source to fetch the required GCP policies for CAST AI integration. No specific configuration is needed for basic usage.

```terraform
data "castai_gke_user_policies" "gke" {}

```

--------------------------------

### Configure AWS Edge Location with Terraform

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/edge_location.md

Examples demonstrating how to define an AWS edge location resource with default, custom, or no addons.

```terraform
# AWS Edge Location with default addons (nvidia-gpu-operator installed by default when addons is omitted)
resource "castai_edge_location" "aws_example" {
  organization_id = "your-org-id"
  cluster_id      = castai_omni_cluster.example.id
  name            = "aws-edge-us-east"
  description     = "AWS edge location in us-east-1"
  region          = "us-east-1"

  zones = [
    {
      id   = "us-east-1a"
      name = "us-east-1a"
    },
    {
      id   = "us-east-1b"
      name = "us-east-1b"
    }
  ]

  aws = {
    account_id           = "123456789012"
    access_key_id_wo     = "AKIAIOSFODNN7EXAMPLE"
    secret_access_key_wo = "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY"
    vpc_id               = "vpc-12345678"
    security_group_id    = "sg-12345678"
    subnet_ids = {
      "us-east-1a" = "subnet-12345678"
      "us-east-1b" = "subnet-87654321"
    }
    name_tag = "castai-edge-location"
  }
}

# AWS Edge Location with custom addons
resource "castai_edge_location" "aws_example_custom_addons" {
  organization_id = "your-org-id"
  cluster_id      = castai_omni_cluster.example.id
  name            = "aws-edge-us-east-gpu"
  region          = "us-east-1"

  addons = [
    {
      name = "nvidia-gpu-operator"
      values = jsonencode({
        driver = {
          enabled = true
        }
      })
    },
    {
      name = "nvidia-network-operator"
    }
  ]

  aws = {
    account_id        = "123456789012"
    vpc_id            = "vpc-12345678"
    security_group_id = "sg-12345678"
    subnet_ids = {
      "us-east-1a" = "subnet-12345678"
    }
  }
}

# AWS Edge Location with no addons
resource "castai_edge_location" "aws_example_no_addons" {
  organization_id = "your-org-id"
  cluster_id      = castai_omni_cluster.example.id
  name            = "aws-edge-us-east-minimal"
  region          = "us-east-1"

  addons = []

  aws = {
    account_id        = "123456789012"
    vpc_id            = "vpc-12345678"
    security_group_id = "sg-12345678"
    subnet_ids = {
      "us-east-1a" = "subnet-12345678"
    }
  }
}
```

--------------------------------

### Recommended: Use presets for standard metrics

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/workload_custom_metrics_data_source.md

Use presets for standard metrics managed by CAST AI, which are kept up-to-date automatically. This example configures a Prometheus data source with the 'jvm' preset.

```terraform
resource "castai_workload_custom_metrics_data_source" "prometheus" {
  cluster_id = castai_eks_cluster.this.id

  name = "my-prometheus"

  prometheus {
    url     = "http://prometheus-server.monitoring.svc.cluster.local:9090"
    timeout = "30s"

    presets = ["jvm"]
  }
}
```

--------------------------------

### Azure Exported Reservations CSV Structure

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/reservations.md

Provides an example of the CSV format for reservations exported from Azure, including specific Azure-related fields.

```csv
Name,Reservation Id,Reservation order Id,Status,Expiration date,Purchase date,Term,Scope,Scope subscription,Scope resource group,Type,Product name,Region,Quantity,Utilization % 1 Day,Utilization % 7 Day,Utilization % 30 Day,Deep link to reservation
VM_RI_01-01-2023_01-01,3b3de39c-bc44-4d69-be2d-69527dfe9958,630226bb-5170-4b95-90b0-f222757130c1,Succeeded,2050-01-01T00:00:00Z,2023-01-11T00:00:00Z,P3Y,Single subscription,8faa0959-093b-4612-8686-a996ac19db00,All resource groups,VirtualMachines,Standard_D32as_v4,eastus,3,100,100,100,https://portal.azure.com#resource/providers/microsoft.capacity/reservationOrders/59791a62-264b-4b9f-aa3a-5eeb761e4583/reservations/883afd52-54c8-4bc6-a0f2-ccbaf7b84bda/overview
VM_RI_01-01-2023_01-02,3b3de39c-bc44-4d69-be2d-69527dfe9958,630226bb-5170-4b95-90b0-f222757130c1,Succeeded,2050-01-01T00:00:00Z,2023-01-11T00:00:01Z,P3Y,Single subscription,8faa0959-093b-4612-8686-a996ac19db00,All resource groups,VirtualMachines,Standard_D32as_v4,eastus,2,100,100,100,https://portal.azure.com#resource/providers/microsoft.capacity/reservationOrders/59791a62-264b-4b9f-aa3a-5eeb761e4583/reservations/25b95bdb-b78b-4973-a60c-71e70f158eca/overview
VM_RI_01-01-2023_01-03,3b3de39c-bc44-4d69-be2d-69527dfe9958,630226bb-5170-4b95-90b0-f222757130c1,Succeeded,2050-01-01T00:00:00Z,2023-01-11T00:00:02Z,P3Y,Single subscription,8faa0959-093b-4612-8686-a996ac19db00,All resource groups,VirtualMachines,Standard_D32as_v4,eastus,1,100,100,100,https://portal.azure.com#resource/providers/microsoft.capacity/reservationOrders/59791a62-264b-4b9f-aa3a-5eeb761e4583/reservations/1745741b-f3c6-46a9-ad16-b93775a1bc38/overview
```

--------------------------------

### Granting EC2 and Pricing Read Permissions to Kentroller

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_kent/README.MD

This example demonstrates how to create a dedicated IAM role with specific read permissions for EC2 inventory, instance offerings, spot-price, and pricing lookups, and associate it with the `castai-kentroller` ServiceAccount.

```terraform
resource "aws_iam_role" "castai_kentroller" {
  name = "${var.cluster_name}-castai-kentroller"

  assume_role_policy = jsonencode({
    Version = "2012-10-17"
    Statement = [
      {
        Action = "sts:AssumeRole"
        Effect = "Allow"
        Principal = {
          Federated = "arn:aws:iam::${data.aws_caller_identity.current.account_id}:oidc-provider/${data.aws_eks_cluster.castai.identity.0.oidc.0.issuer}"
        }
        Condition = {
          "oidc.eks.amazonaws.com/TokenEx": {
            "${data.aws_eks_cluster.castai.identity.0.oidc.0.issuer}": "system:serviceaccount:castai-agent:castai-kentroller"
          }
        }
      }
    ]
  })
}

resource "aws_iam_role_policy" "castai_kentroller" {
  name = "castai-kentroller"
  role = aws_iam_role.castai_kentroller.id

  policy = jsonencode({
    Version = "2012-10-17"
    Statement = [
      {
        Action = [
          "ec2:DescribeInstanceTypes",
          "ec2:DescribeInstances",
          "ec2:DescribeImages",
          "ec2:DescribeSubnets",
          "ec2:DescribeSecurityGroups",
          "ec2:DescribeVpcs",
          "ec2:DescribeAddresses",
          "ec2:DescribeNetworkInterfaces",
          "ec2:DescribeInstanceAttribute",
          "ec2:DescribeInstanceTopology",
          "ec2:DescribeInstanceStatus",
          "ec2:DescribeEbsVolumes",
          "ec2:DescribeVolumesModifications",
          "ec2:DescribeSnapshots",
          "ec2:DescribeTags",
          "ec2:DescribeRegions",
          "ec2:DescribeAvailabilityZones",
          "ec2:DescribeHostTenancy"
        ]
        Effect   = "Allow"
        Resource = "*"
      },
      {
        Action   = "pricing:GetProducts"
        Effect   = "Allow"
        Resource = "*"
      },
      {
        Action   = "eks:DescribeCluster"
        Effect   = "Allow"
        Resource = "${data.aws_eks_cluster.castai.arn}"
      }
    ]
  })
}

resource "aws_eks_pod_identity_association" "castai_kentroller" {
  cluster_id     = data.aws_eks_cluster.castai.id
  namespace      = "castai-agent"
  service_account_name = "castai-kentroller"
  role_arn       = aws_iam_role.castai_kentroller.arn
}

```

--------------------------------

### Fetch CAST AI Runtime Rules

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_with_security_runtime_rules/README.MD

Execute the provided shell script to fetch CAST AI runtime rules. This script requires 'jq' to be installed. You can fetch all rules or only those that are already enabled.

```bash
# Fetch all rules
./fetch_castai_runtime_rules.sh

# Fetch only already enabled rules
./fetch_castai_runtime_rules.sh --enabled-only
```

--------------------------------

### Deploy CAST AI Spot Handler on AKS

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/aks/aks_cluster_gitops/README.md

Installs the spot handler component. This command configures it to use the cluster's metadata and sets the provider to Azure. Ensure the `CAST_CONFIG_CLUSTERID` is correctly set.

```bash
helm upgrade -i castai-spot-handler castai-helm/castai-spot-handler -n castai-agent \
--set "envFrom[0].configMapRef.name=$CAST_CONFIG_CLUSTERID" \
--set castai.provider=azure
```

--------------------------------

### Deploy CAST AI Cluster Controller using Helm

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_gitops/README.md

Installs or upgrades the castai-cluster-controller Helm chart. This component requires specific environment variables for configuration and enabling autoscaling.

```bash
helm upgrade -i cluster-controller castai-helm/castai-cluster-controller -n castai-agent \
  --set autoscaling.enabled=true \
  --set "envFrom[0].secretRef.name=$CAST_SECRET_APIKEY" \
  --set "envFrom[1].configMapRef.name=$CAST_CONFIG_CLUSTERID"
```

--------------------------------

### Deploy CAST AI Agent using Helm

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_gitops/README.md

Installs or upgrades the castai-agent Helm chart. This is a mandatory component for cluster onboarding and requires the CAST AI API key and cluster ID.

```bash
CASTAI_API_KEY="<Replace cluster_token>"
CAST_CONFIG_CLUSTERID="castai-agent-metadata"
CAST_SECRET_APIKEY="castai-agent"

helm upgrade -i castai-agent castai-helm/castai-agent -n castai-agent --create-namespace \
  --set apiKey="$CASTAI_API_KEY" \
  --set provider=eks \
  --set createNamespace=false \
  --set metadataStore.enabled=true
```

--------------------------------

### Importing an Existing Scaling Policy Order

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/workload_scaling_policy_order.md

This example shows how to import an existing workload scaling policy order into your Terraform state. Ensure you have the correct cluster ID and adjust the import configuration as needed.

```terraform
import {
 to = castai_workload_scaling_policy_order.custom
 id = var.castai_gke_cluster_id
}
```

--------------------------------

### EKS Settings Data Source Example

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/data-sources/eks_settings.md

Use this data source to fetch EKS settings, including IAM policies, for a specified cluster. Ensure you have the necessary AWS credentials and cluster information configured.

```terraform
data "castai_eks_settings" "current" {
  account_id = data.aws_caller_identity.current.account_id
  region     = var.cluster_region
  cluster    = var.cluster_name
  vpc        = var.vpc
}
```

--------------------------------

### Import CAST AI Runtime Rules Script

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/aks/aks_cluster_with_security_runtime_rules/README.MD

This shell script fetches CAST AI runtime rules and generates Terraform resources. It requires `jq` to be installed. Set `CASTAI_API_KEY` and optionally `CASTAI_BASE_URL` environment variables before running.

```shell
export CASTAI_API_KEY="your-api-key-here"
export CASTAI_BASE_URL="https://api.cast.ai"  # Optional, default is https://api.cast.ai

# Fetch all rules
./fetch_castai_runtime_rules.sh

# Fetch only already enabled rules
./fetch_castai_runtime_rules.sh --enabled-only
```

--------------------------------

### Deploy CAST AI Agent on AKS

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/aks/aks_cluster_gitops/README.md

Installs the core CAST AI agent using Helm. Ensure you replace the placeholder for CASTAI_API_KEY with your actual cluster token. This component is mandatory for cluster onboarding.

```bash
CASTAI_API_KEY="<Replace cluster_token>"
CAST_CONFIG_CLUSTERID="castai-agent-metadata"
CAST_SECRET_APIKEY="castai-agent"

helm upgrade -i castai-agent castai-helm/castai-agent -n castai-agent --create-namespace \
  --set apiKey="$CASTAI_API_KEY" \
  --set provider=aks \
  --set createNamespace=false \
  --set metadataStore.enabled=true
```

--------------------------------

### Deploy CAST AI Pod Pinner on AKS

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/aks/aks_cluster_gitops/README.md

Installs the pod pinner component. Note that `replicaCount` is set to 0, indicating it might be disabled by default or configured for manual activation. It requires references to secrets and config maps.

```bash
helm upgrade -i castai-pod-pinner castai-helm/castai-pod-pinner -n castai-agent \
--set "envFrom[0].secretRef.name=$CAST_SECRET_APIKEY" \
--set "envFrom[1].configMapRef.name=$CAST_CONFIG_CLUSTERID" \
--set replicaCount=0
```

--------------------------------

### Initialize and Apply Terraform Configuration

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_existing_with_ai_optimizer/README.MD

Initialize Terraform and apply the configuration to set up CAST AI resources, NVIDIA device plugin, and model deployments. Use `terraform destroy` to remove created resources.

```bash
terraform init
terraform apply
terraform destroy
```

--------------------------------

### Generate import plan

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/workload_scaling_policy.md

Execute the plan command to generate configuration for imported resources.

```shell
terraform plan -out=import.plan -var-file=tf.vars -generate-config-out=generated.tf
```

--------------------------------

### Apply Terraform Configuration

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/aks/aks_cluster/README.MD

Execute this command to create the resources defined in your Terraform configuration.

```bash
terraform apply
```

--------------------------------

### Initialize and Apply Terraform Configuration

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_kent/README.MD

Initialize the Terraform working directory and apply the configuration to deploy the EKS cluster stack. Ensure you have filled in the required variables in `terraform.tfvars`.

```sh
cp terraform.tfvars.example terraform.tfvars
# fill in cluster_name, cluster_region, castai_api_token
terraform init
terraform apply
```

--------------------------------

### Install CAST AI in read-only mode

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/gke/gke_cluster_gitops_umbrella_chart/README.md

Installs the CAST AI umbrella Helm chart in read-only mode on a GKE cluster. This mode allows CAST AI to observe the cluster without making any changes to workloads or nodes. Requires a CAST AI member API key and Helm.

```sh
helm upgrade -i castai castai-helm/castai -n castai-agent --create-namespace \
  --set global.castai.apiKey="<your-castai-api-key>" \
  --set global.castai.provider="gke" \
  --set tags.readonly=true
```

--------------------------------

### Enable Debug Logging for Terraform

Source: https://github.com/castai/terraform-provider-castai/blob/master/README.md

To get more detailed logs during Terraform operations, set the `TF_LOG` environment variable to `DEBUG`.

```bash
TF_LOG=DEBUG terraform plan
```

--------------------------------

### Manage Cloud Commitments with Terraform

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/commitment.md

Examples of defining various cloud provider commitments including GCP, AWS, and Azure resources.

```terraform
# GCP committed use discount (resource CUD).
resource "castai_commitment" "gcp_cud" {
  name               = "prod-cud-us-central1"
  cloud              = "GCP"
  region             = "us-central1"
  type               = "RESOURCE_CUD"
  start_time         = "2026-01-01T00:00:00Z"
  end_time           = "2027-01-01T00:00:00Z"
  autoscaling_status = "ACTIVE"
  allowed_usage      = 1.0

  gcp_resource_cud_details = {
    cud_id    = "123456789"
    plan      = "TWELVE_MONTH"
    type      = "GENERAL_PURPOSE_E2"
    cpu       = 32
    memory_mb = 131072
    status    = "ACTIVE"
  }
}

# AWS reserved instances.
resource "castai_commitment" "aws_ri" {
  name       = "prod-ri-us-east-1"
  cloud      = "AWS"
  region     = "us-east-1"
  type       = "RESERVED_INSTANCE"
  start_time = "2026-01-01T00:00:00Z"
  end_time   = "2027-01-01T00:00:00Z"

  aws_reserved_instances_details = {
    id             = "abcdef01-2345-6789-abcd-ef0123456789"
    scope          = "Region"
    instance_type  = "m5.xlarge"
    instance_count = 10
    state          = "active"
  }
}

# Azure reservation.
resource "castai_commitment" "azure_reservation" {
  name       = "prod-reservation-eastus"
  cloud      = "AZURE"
  region     = "eastus"
  type       = "RESERVED_INSTANCE"
  start_time = "2026-01-01T00:00:00Z"
  end_time   = "2029-01-01T00:00:00Z"

  azure_reservation_details = {
    id                   = "abcdef01-2345-6789-abcd-ef0123456789"
    plan                 = "THREE_YEAR"
    status               = "Succeeded"
    scope                = "Shared"
    instance_type        = "Standard_D4s_v3"
    count                = 5
    instance_flexibility = "ON"
  }
}

# GCP capacity reservation (on-demand).
resource "castai_commitment" "gcp_capacity_reservation" {
  name       = "gpu-reservation-us-central1"
  cloud      = "GCP"
  region     = "us-central1"
  type       = "ON_DEMAND_CAPACITY_RESERVATION"
  start_time = "2026-01-01T00:00:00Z"

  gcp_capacity_reservation_details = {
    id                   = "my-reservation"
    self_link            = "https://www.googleapis.com/compute/v1/projects/my-project/zones/us-central1-a/reservations/my-reservation"
    project_id           = "my-project"
    zone                 = "us-central1-a"
    instance_type        = "a2-highgpu-1g"
    total_instance_count = 4
    state                = "READY"

    accelerators = [
      {
        accelerator_type  = "https://www.googleapis.com/compute/v1/projects/my-project/zones/us-central1-a/acceleratorTypes/nvidia-tesla-a100"
        accelerator_count = 1
      }
    ]
  }
}

# AWS on-demand capacity reservation (ODCR).
resource "castai_commitment" "aws_odcr" {
  name               = "prod-odcr-us-east-1"
  cloud              = "AWS"
  region             = "us-east-1"
  type               = "ON_DEMAND_CAPACITY_RESERVATION"
  start_time         = "2026-01-01T00:00:00Z"
  autoscaling_status = "ACTIVE"
  allowed_usage      = 1.0

  aws_odcr_details = {
    id                       = "cr-abcdef01234567890"
    availability_zone        = "us-east-1a"
    instance_type            = "m5.xlarge"
    instance_platform        = "Linux/UNIX"
    tenancy                  = "default"
    total_instance_count     = 10
    available_instance_count = 10
    state                    = "active"
    end_date_type            = "unlimited"
    instance_match_criteria  = "open"
    interruptible            = false
  }
}
```

--------------------------------

### Create Organization Workspace and Resources

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/gke/gke_clusters_with_workspaces/README.md

Create a new Terraform workspace for the organization and apply the initial configuration using the organization's variable file.

```bash
terraform workspace new org-workspace
terraform plan -var-file=tf.vars
terraform apply -var-file=tf.vars
```

--------------------------------

### Plan import for multiple policies

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/workload_scaling_policy.md

Generate an import plan for multiple policies without generating new configuration files.

```shell
terraform plan -out=import.plan -var-file=tf.vars
```

--------------------------------

### Define resource for CLI import

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/workload_scaling_policy.md

Establish the resource block in configuration before running the import command.

```hcl
resource "castai_workload_scaling_policy" "services" {
  # ...
}
```

--------------------------------

### Terraform Lifecycle Commands

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_readonly_kvisor/README.MD

Standard Terraform commands to initialize, apply, and destroy the infrastructure configuration.

```bash
terraform init
```

```bash
terraform apply
```

```bash
terraform destroy
```

--------------------------------

### Plan Terraform Deployment

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/anywhere/README.md

Previews the resources Terraform will create or modify.

```sh
terraform plan
```

--------------------------------

### Azure Commitments Configuration

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/commitments.md

Example of configuring Azure Reserved Instances with specific matchers, allowed usage, and cluster assignments. Auto-assignment is enabled by default.

```terraform
resource "castai_commitments" "azure_test" {
  azure_reservations_csv = file("./reservations.csv")
  commitment_configs {
    matcher {
      region = "eastus"
      type   = "Standard_D32as_v4"
      name   = "test-res-1"
    }
    prioritization   = false
    allowed_usage    = 0.9
    status           = "Active"
    scaling_strategy = "Default"

    assignments {
      cluster_id = "cluster-id-3"
    }
    assignments {
      cluster_id = "cluster-id-4"
    }
  }
}
```

--------------------------------

### Add castai-helm repository

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/gke/gke_cluster_gitops_umbrella_chart/README.md

Adds the castai-helm repository to your Helm configuration and updates the repository list. This is a prerequisite for installing the CAST AI Helm chart.

```sh
helm repo add castai-helm https://castai.github.io/helm-charts
helm repo update
```

--------------------------------

### Enable Extended Permissions in Terraform

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/castware-operator/README.md

Set the `extended_permissions` variable to `true` in your Terraform configuration to enable extended permissions, which is required for installing the `cluster-controller` component.

```hcl
extended_permissions = true
```

--------------------------------

### Initialize and apply Terraform for Full mode

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/gke/gke_cluster_gitops_umbrella_chart/README.md

Initializes the Terraform configuration and applies it to create a GCP service account with necessary IAM permissions for node provisioning and to register the cluster with CAST AI. This step is required before upgrading to Full mode.

```sh
terraform init
terraform apply
```

--------------------------------

### Expected CAST AI Stream Connection Log Output

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_kent/README.MD

This is an example of successful log messages indicating a healthy stream connection and registration with CAST AI.

```log
msg="getStream: stream registered confirmation received, stream is ready" cluster_id=<uuid>
msg="streaming messages..." cluster_id=<uuid>
msg="stream connected, syncing all schedules to Kent" cluster_id=<uuid>
```

--------------------------------

### Run Unit Tests for CastAI Terraform Provider

Source: https://github.com/castai/terraform-provider-castai/blob/master/README.md

Execute the unit tests for the provider using the `make test` command.

```bash
make test
```

--------------------------------

### Get Service Account Unique IDs

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/gke/gke_castai_iam/README.md

Run this script to obtain the unique IDs for service accounts, which are necessary for configuring IAM policies with custom conditions.

```bash
PROJECT_ID=<PLACEHOLDER> LOCATION=<PLACEHOLDER> CLUSTER_NAME=<PLACEHOLDER> ./script.sh
```

--------------------------------

### Configure Terraform for Local Provider Development

Source: https://github.com/castai/terraform-provider-castai/blob/master/README.md

Set up the `~/.terraformrc` file to use a locally built provider. Replace `<path-to-terraform-provider-castai-repository>` with the actual path to your cloned repository.

```terraform
provider_installation {
  dev_overrides {
    "castai/castai" = "<path-to-terraform-provider-castai-repository>"
  }
  direct {}
}
```

--------------------------------

### Deploy CAST AI Workload Autoscaler using Helm

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_gitops/README.md

Installs or upgrades the castai-workload-autoscaler Helm chart. This component requires specific secret and configMap references for its configuration.

```bash
helm upgrade -i castai-workload-autoscaler castai-helm/castai-workload-autoscaler -n castai-agent \
--set "envFrom[0].secretRef.name=$CAST_SECRET_APIKEY" \
--set "envFrom[1].configMapRef.name=$CAST_CONFIG_CLUSTERID" \

```

--------------------------------

### Import Node Configuration

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/node_configuration.md

Execute the import command to attach an existing node configuration to the defined Terraform resource.

```shell
$ terraform import castai_node_configuration.default <cluster_id>/default
```

```shell
$ terraform import 'module.castai-eks-cluster.castai_node_configuration.this["default"]' <cluster_id>/default
```

--------------------------------

### Retrieve CAST AI Cluster ID

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_kent/README.MD

Get the CAST AI cluster ID from the agent's ConfigMap, which is used for identification within the CAST AI platform.

```sh
kubectl -n castai-agent get cm castai-agent-metadata -o jsonpath='{.data.CLUSTER_ID}'
```

--------------------------------

### Configure S3 Model Registry

Source: https://github.com/castai/terraform-provider-castai/blob/master/docs/resources/ai_optimizer_model_registry.md

Example of how to configure an AWS S3 bucket as the AI optimizer model registry. Ensure the `registry_credentials` variable is properly set with JSON-encoded credentials.

```terraform
resource "castai_ai_optimizer_model_registry" "example" {
  provider_type = "S3"
  credentials   = var.registry_credentials

  s3 {
    bucket = "my-company-model-registry"
    region = "us-east-1"
    prefix = "models/"
  }
}
```

--------------------------------

### Enable CAST AI Live Migrations (CLM)

Source: https://github.com/castai/terraform-provider-castai/blob/master/examples/eks/eks_cluster_kent/README.MD

Configure the `kent` subchart to enable live migrations by setting the CLM controller replica count.

```hcl
kent = {
  enabled = true
  castai-live = {
    controller = { replicaCount = 2 }
  }
}
```