feat(eks-example): expose agent_values and cluster_controller_values for Helm customization

Adds agent_values and cluster_controller_values variables to the
eks_cluster_existing example, enabling users to pass arbitrary Helm
values (e.g. additionalEnv) to the castai-agent and cluster-controller
charts without modifying castai.tf directly.

Also fixes stale variable references in README and terraform.tfvars.example.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: Eric Kufta

examples/eks/eks_cluster_existing/README.MD

@@ -9,17 +9,18 @@ Example configuration should be analysed in the following order:
 
 # Usage
 1. Copy `terraform.tfvars.example` to `terraform.tfvars`
-2. Update `terraform.tfvars` file with your cluster name, cluster region, vpc_id, cluster_security_group_id, node_security_group_id, subnets and CAST AI API token.
+2. Update `terraform.tfvars` file with your cluster name, cluster region, and CAST AI API token.
 
 | Variable | Description |
 | --- | --- |
-| cluster_name                = "" | Name of cluster |
-| cluster_region              = "" | Name of region of cluster |
-| castai_api_token            = "" | Cast api token |
-| vpc_id                      = "" | Virtual Private Cloud(VPC) id |
-| cluster_security_group_id   = "" | Cluster security group id |
-| node_security_group_id      = "" | Node security group id |
-| subnets                     = ["", ""] | Public subnets of cluster |
+| `cluster_name` | Name of the EKS cluster |
+| `cluster_region` | AWS region where the cluster runs |
+| `castai_api_token` | CAST AI API token from console.cast.ai → API Access keys |
+| `profile` | AWS CLI profile to use (default: `"default"`) |
+| `delete_nodes_on_disconnect` | Delete CAST AI-provisioned nodes on disconnect (default: `true`; set to `false` for production) |
+| `tags` | Optional map of tags applied to new nodes |
+| `agent_values` | Optional list of YAML Helm values for castai-agent. Example: `[<<-EOF\n  additionalEnv:\n    MY_VAR: "value"\nEOF\n]` — see [Customizing Helm values](#customizing-helm-values-additionalenv-and-more) |
+| `cluster_controller_values` | Optional list of YAML Helm values for castai-cluster-controller, same format as `agent_values` |
 
 3. Initialize Terraform. Under example root folder run:
 ```
@@ -44,4 +45,22 @@ AWS CLI profile is already set to default, override if only required.
 > used by CAST AI has to be present. Example of entry can be found [here](https://github.com/castai/terraform-provider-castai/blob/157babd57b0977f499eb162e9bee27bee51d292a/examples/eks/eks_cluster_autoscaler_polices/eks.tf#L28-L38).
 
 
+# Customizing Helm values (additionalEnv and more)
+
+The `agent_values` and `cluster_controller_values` variables accept lists of YAML-formatted Helm values that are merged into the respective CAST AI Helm chart releases. This is the recommended way to configure environment variables, proxy settings, resource limits, or any other chart-level option.
+
+Example `terraform.tfvars` entry to set environment variables on the agent:
+
+```hcl
+agent_values = [<<-EOF
+  additionalEnv:
+    HTTP_PROXY: "http://proxy.example.com:8080"
+    HTTPS_PROXY: "http://proxy.example.com:8080"
+    NO_PROXY: "169.254.169.254,localhost"
+EOF
+]
+```
+
+Use `cluster_controller_values` for the same pattern applied to the cluster-controller component.
+
 Please refer to this guide if you run into any issues https://docs.cast.ai/docs/terraform-troubleshooting

examples/eks/eks_cluster_existing/castai.tf

@@ -190,6 +190,22 @@ module "castai_eks_cluster" {
   # Installs network monitor
   install_egressd = true
 
+  # Optional: pass additional Helm values to castai-agent and cluster-controller.
+  # This is useful for setting environment variables (additionalEnv), resource limits,
+  # pod annotations, proxy settings, etc.
+  #
+  # Example terraform.tfvars entry:
+  #
+  #   agent_values = [<<-EOF
+  #     additionalEnv:
+  #       MY_ENV_VAR: "my-value"
+  #       HTTP_PROXY: "http://proxy.example.com:8080"
+  #   EOF
+  #   ]
+  #
+  agent_values              = var.agent_values
+  cluster_controller_values = var.cluster_controller_values
+
   # depends_on helps Terraform with creating proper dependencies graph in case of resource creation and in this case destroy.
   # module "castai-eks-cluster" has to be destroyed before module "castai-eks-role-iam".
   depends_on = [module.castai_eks_role_iam]

examples/eks/eks_cluster_existing/terraform.tfvars.example

@@ -1,7 +1,4 @@
-cluster_name                = ""
-cluster_region              = ""
-castai_api_token            = ""
-grpc_url                    = ""
-vpc_id                      = ""
-subnets                     = ["", ""]
-profile                     = "default" # default aws cli profile is used, override if needed.
+cluster_name     = ""
+cluster_region   = ""
+castai_api_token = ""
+profile          = "default" # default aws cli profile is used, override if needed.

examples/eks/eks_cluster_existing/variables.tf

@@ -44,3 +44,15 @@ variable "tags" {
   description = "Optional tags for new cluster nodes. This parameter applies only to new nodes - tags for old nodes are not reconciled."
   default     = {}
 }
+
+variable "agent_values" {
+  type        = list(string)
+  description = "List of YAML-formatted Helm values for the castai-agent chart. Use this to set additionalEnv, resource limits, pod annotations, etc."
+  default     = []
+}
+
+variable "cluster_controller_values" {
+  type        = list(string)
+  description = "List of YAML-formatted Helm values for the castai-cluster-controller chart. Use this to set additionalEnv, resource limits, pod annotations, etc."
+  default     = []
+}