feat: Add support for Pod Identity assocation on Karpenter sub-module (#3031)

Co-authored-by: Bryant Biggs <bryantbiggs@gmail.com>

examples/karpenter/README.md

@@ -1,6 +1,6 @@
 # Karpenter Example
 
-Configuration in this directory creates an AWS EKS cluster with [Karpenter](https://karpenter.sh/) provisioned for managing compute resource scaling. In the example provided, Karpenter is running on EKS Fargate yet Karpenter is providing compute in the form of EC2 instances.
+Configuration in this directory creates an AWS EKS cluster with [Karpenter](https://karpenter.sh/) provisioned for managing compute resource scaling. In the example provided, Karpenter is provisioned on top of an EKS Managed Node Group.
 
 ## Usage
 
@@ -22,10 +22,47 @@ aws eks --region eu-west-1 update-kubeconfig --name ex-karpenter
 kubectl scale deployment inflate --replicas 5
 
 # You can watch Karpenter's controller logs with
-kubectl logs -f -n karpenter -l app.kubernetes.io/name=karpenter -c controller
+kubectl logs -f -n kube-system -l app.kubernetes.io/name=karpenter -c controller
 ```
 
-You should see a new node named `karpenter.sh/provisioner-name/default` eventually come up in the console; this was provisioned by Karpenter in response to the scaled deployment above.
+Validate if the Amazon EKS Addons Pods are running in the Managed Node Group and the `inflate` application Pods are running on Karpenter provisioned Nodes.
+
+```bash
+kubectl get nodes -L karpenter.sh/registered
+```
+
+```text
+NAME                                        STATUS   ROLES    AGE    VERSION               REGISTERED
+ip-10-0-16-155.eu-west-1.compute.internal   Ready    <none>   100s   v1.29.3-eks-ae9a62a   true
+ip-10-0-3-23.eu-west-1.compute.internal     Ready    <none>   6m1s   v1.29.3-eks-ae9a62a
+ip-10-0-41-2.eu-west-1.compute.internal     Ready    <none>   6m3s   v1.29.3-eks-ae9a62a
+```
+
+```sh
+kubectl get pods -A -o custom-columns=NAME:.metadata.name,NODE:.spec.nodeName
+```
+
+```text
+NAME                           NODE
+inflate-75d744d4c6-nqwz8       ip-10-0-16-155.eu-west-1.compute.internal
+inflate-75d744d4c6-nrqnn       ip-10-0-16-155.eu-west-1.compute.internal
+inflate-75d744d4c6-sp4dx       ip-10-0-16-155.eu-west-1.compute.internal
+inflate-75d744d4c6-xqzd9       ip-10-0-16-155.eu-west-1.compute.internal
+inflate-75d744d4c6-xr6p5       ip-10-0-16-155.eu-west-1.compute.internal
+aws-node-mnn7r                 ip-10-0-3-23.eu-west-1.compute.internal
+aws-node-rkmvm                 ip-10-0-16-155.eu-west-1.compute.internal
+aws-node-s4slh                 ip-10-0-41-2.eu-west-1.compute.internal
+coredns-68bd859788-7rcfq       ip-10-0-3-23.eu-west-1.compute.internal
+coredns-68bd859788-l78hw       ip-10-0-41-2.eu-west-1.compute.internal
+eks-pod-identity-agent-gbx8l   ip-10-0-41-2.eu-west-1.compute.internal
+eks-pod-identity-agent-s7vt7   ip-10-0-16-155.eu-west-1.compute.internal
+eks-pod-identity-agent-xwgqw   ip-10-0-3-23.eu-west-1.compute.internal
+karpenter-79f59bdfdc-9q5ff     ip-10-0-41-2.eu-west-1.compute.internal
+karpenter-79f59bdfdc-cxvhr     ip-10-0-3-23.eu-west-1.compute.internal
+kube-proxy-7crbl               ip-10-0-41-2.eu-west-1.compute.internal
+kube-proxy-jtzds               ip-10-0-16-155.eu-west-1.compute.internal
+kube-proxy-sm42c               ip-10-0-3-23.eu-west-1.compute.internal
+```
 
 ### Tear Down & Clean-Up
 

examples/karpenter/main.tf

@@ -41,9 +41,8 @@ data "aws_ecrpublic_authorization_token" "token" {
 }
 
 locals {
-  name            = "ex-${replace(basename(path.cwd), "_", "-")}"
+  name   = "ex-${basename(path.cwd)}"
-  cluster_version = "1.29"
+  region = "eu-west-1"
-  region          = "eu-west-1"
 
   vpc_cidr = "10.0.0.0/16"
   azs      = slice(data.aws_availability_zones.available.names, 0, 3)
@@ -62,62 +61,42 @@ locals {
 module "eks" {
   source = "../.."
 
-  cluster_name                   = local.name
+  cluster_name    = local.name
-  cluster_version                = local.cluster_version
+  cluster_version = "1.29"
-  cluster_endpoint_public_access = true
 
   # Gives Terraform identity admin access to cluster which will
   # allow deploying resources (Karpenter) into the cluster
   enable_cluster_creator_admin_permissions = true
+  cluster_endpoint_public_access           = true
 
   cluster_addons = {
-    coredns = {
+    coredns                = {}
-      configuration_values = jsonencode({
+    eks-pod-identity-agent = {}
-        computeType = "Fargate"
+    kube-proxy             = {}
-        # Ensure that we fully utilize the minimum amount of resources that are supplied by
+    vpc-cni                = {}
-        # Fargate https://docs.aws.amazon.com/eks/latest/userguide/fargate-pod-configuration.html
-        # Fargate adds 256 MB to each pod's memory reservation for the required Kubernetes
-        # components (kubelet, kube-proxy, and containerd). Fargate rounds up to the following
-        # compute configuration that most closely matches the sum of vCPU and memory requests in
-        # order to ensure pods always have the resources that they need to run.
-        resources = {
-          limits = {
-            cpu = "0.25"
-            # We are targeting the smallest Task size of 512Mb, so we subtract 256Mb from the
-            # request/limit to ensure we can fit within that task
-            memory = "256M"
-          }
-          requests = {
-            cpu = "0.25"
-            # We are targeting the smallest Task size of 512Mb, so we subtract 256Mb from the
-            # request/limit to ensure we can fit within that task
-            memory = "256M"
-          }
-        }
-      })
-    }
-    kube-proxy = {}
-    vpc-cni    = {}
   }
 
   vpc_id                   = module.vpc.vpc_id
   subnet_ids               = module.vpc.private_subnets
   control_plane_subnet_ids = module.vpc.intra_subnets
 
-  # Fargate profiles use the cluster primary security group so these are not utilized
+  eks_managed_node_groups = {
-  create_cluster_security_group = false
-  create_node_security_group    = false
-
-  fargate_profiles = {
     karpenter = {
-      selectors = [
+      instance_types = ["m5.large"]
-        { namespace = "karpenter" }
+
-      ]
+      min_size     = 2
-    }
+      max_size     = 3
-    kube-system = {
+      desired_size = 2
-      selectors = [
+
-        { namespace = "kube-system" }
+      taints = {
-      ]
+        # This Taint aims to keep just EKS Addons and Karpenter running on this MNG
+        # The pods that do not tolerate this taint should run on nodes created by Karpenter
+        addons = {
+          key    = "CriticalAddonsOnly"
+          value  = "true"
+          effect = "NO_SCHEDULE"
+        },
+      }
     }
   }
 
@@ -138,9 +117,8 @@ module "karpenter" {
 
   cluster_name = module.eks.cluster_name
 
-  # EKS Fargate currently does not support Pod Identity
+  enable_pod_identity             = true
-  enable_irsa            = true
+  create_pod_identity_association = true
-  irsa_oidc_provider_arn = module.eks.oidc_provider_arn
 
   # Used to attach additional IAM policies to the Karpenter node IAM role
   node_iam_role_additional_policies = {
@@ -162,14 +140,13 @@ module "karpenter_disabled" {
 ################################################################################
 
 resource "helm_release" "karpenter" {
-  namespace           = "karpenter"
+  namespace           = "kube-system"
-  create_namespace    = true
   name                = "karpenter"
   repository          = "oci://public.ecr.aws/karpenter"
   repository_username = data.aws_ecrpublic_authorization_token.token.user_name
   repository_password = data.aws_ecrpublic_authorization_token.token.password
   chart               = "karpenter"
-  version             = "0.35.1"
+  version             = "0.36.1"
   wait                = false
 
   values = [
@@ -178,14 +155,6 @@ resource "helm_release" "karpenter" {
       clusterName: ${module.eks.cluster_name}
       clusterEndpoint: ${module.eks.cluster_endpoint}
       interruptionQueue: ${module.karpenter.queue_name}
-    serviceAccount:
-      annotations:
-        eks.amazonaws.com/role-arn: ${module.karpenter.iam_role_arn}
-    tolerations:
-      - key: 'eks.amazonaws.com/compute-type'
-        operator: Equal
-        value: fargate
-        effect: "NoSchedule"
     EOT
   ]
 }

modules/karpenter/README.md

@@ -8,6 +8,7 @@ Configuration in this directory creates the AWS resources required by Karpenter
 
 In the following example, the Karpenter module will create:
 - An IAM role for use with Pod Identity and a scoped IAM policy for the Karpenter controller
+- A Pod Identity association to grant Karpenter controller access provided by the IAM Role
 - A Node IAM role that Karpenter will use to create an Instance Profile for the nodes to receive IAM permissions
 - An access entry for the Node IAM role to allow nodes to join the cluster
 - SQS queue and EventBridge event rules for Karpenter to utilize for spot termination handling, capacity re-balancing, etc.
@@ -104,6 +105,7 @@ No modules.
 | [aws_cloudwatch_event_rule.this](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/cloudwatch_event_rule) | resource |
 | [aws_cloudwatch_event_target.this](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/cloudwatch_event_target) | resource |
 | [aws_eks_access_entry.node](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/eks_access_entry) | resource |
+| [aws_eks_pod_identity_association.karpenter](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/eks_pod_identity_association) | resource |
 | [aws_iam_instance_profile.this](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/iam_instance_profile) | resource |
 | [aws_iam_policy.controller](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/iam_policy) | resource |
 | [aws_iam_role.controller](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/iam_role) | resource |
@@ -135,6 +137,7 @@ No modules.
 | <a name="input_create_iam_role"></a> [create\_iam\_role](#input\_create\_iam\_role) | Determines whether an IAM role is created | `bool` | `true` | no |
 | <a name="input_create_instance_profile"></a> [create\_instance\_profile](#input\_create\_instance\_profile) | Whether to create an IAM instance profile | `bool` | `false` | no |
 | <a name="input_create_node_iam_role"></a> [create\_node\_iam\_role](#input\_create\_node\_iam\_role) | Determines whether an IAM role is created or to use an existing IAM role | `bool` | `true` | no |
+| <a name="input_create_pod_identity_association"></a> [create\_pod\_identity\_association](#input\_create\_pod\_identity\_association) | Determines whether to create pod identity association | `bool` | `false` | no |
 | <a name="input_enable_irsa"></a> [enable\_irsa](#input\_enable\_irsa) | Determines whether to enable support for IAM role for service accounts | `bool` | `false` | no |
 | <a name="input_enable_pod_identity"></a> [enable\_pod\_identity](#input\_enable\_pod\_identity) | Determines whether to enable support for EKS pod identity | `bool` | `true` | no |
 | <a name="input_enable_spot_termination"></a> [enable\_spot\_termination](#input\_enable\_spot\_termination) | Determines whether to enable native spot termination handling | `bool` | `true` | no |
@@ -153,6 +156,7 @@ No modules.
 | <a name="input_irsa_assume_role_condition_test"></a> [irsa\_assume\_role\_condition\_test](#input\_irsa\_assume\_role\_condition\_test) | Name of the [IAM condition operator](https://docs.aws.amazon.com/IAM/latest/UserGuide/reference_policies_elements_condition_operators.html) to evaluate when assuming the role | `string` | `"StringEquals"` | no |
 | <a name="input_irsa_namespace_service_accounts"></a> [irsa\_namespace\_service\_accounts](#input\_irsa\_namespace\_service\_accounts) | List of `namespace:serviceaccount`pairs to use in trust policy for IAM role for service accounts | `list(string)` | <pre>[<br>  "karpenter:karpenter"<br>]</pre> | no |
 | <a name="input_irsa_oidc_provider_arn"></a> [irsa\_oidc\_provider\_arn](#input\_irsa\_oidc\_provider\_arn) | OIDC provider arn used in trust policy for IAM role for service accounts | `string` | `""` | no |
+| <a name="input_namespace"></a> [namespace](#input\_namespace) | Namespace to associate with the Karpenter Pod Identity | `string` | `"kube-system"` | no |
 | <a name="input_node_iam_role_additional_policies"></a> [node\_iam\_role\_additional\_policies](#input\_node\_iam\_role\_additional\_policies) | Additional policies to be added to the IAM role | `map(string)` | `{}` | no |
 | <a name="input_node_iam_role_arn"></a> [node\_iam\_role\_arn](#input\_node\_iam\_role\_arn) | Existing IAM role ARN for the IAM instance profile. Required if `create_iam_role` is set to `false` | `string` | `null` | no |
 | <a name="input_node_iam_role_attach_cni_policy"></a> [node\_iam\_role\_attach\_cni\_policy](#input\_node\_iam\_role\_attach\_cni\_policy) | Whether to attach the `AmazonEKS_CNI_Policy`/`AmazonEKS_CNI_IPv6_Policy` IAM policy to the IAM IAM role. WARNING: If set `false` the permissions must be assigned to the `aws-node` DaemonSet pods via another method or nodes will not be able to join the cluster | `bool` | `true` | no |
@@ -168,6 +172,7 @@ No modules.
 | <a name="input_queue_managed_sse_enabled"></a> [queue\_managed\_sse\_enabled](#input\_queue\_managed\_sse\_enabled) | Boolean to enable server-side encryption (SSE) of message content with SQS-owned encryption keys | `bool` | `true` | no |
 | <a name="input_queue_name"></a> [queue\_name](#input\_queue\_name) | Name of the SQS queue | `string` | `null` | no |
 | <a name="input_rule_name_prefix"></a> [rule\_name\_prefix](#input\_rule\_name\_prefix) | Prefix used for all event bridge rules | `string` | `"Karpenter"` | no |
+| <a name="input_service_account"></a> [service\_account](#input\_service\_account) | Service account to associate with the Karpenter Pod Identity | `string` | `"karpenter"` | no |
 | <a name="input_tags"></a> [tags](#input\_tags) | A map of tags to add to all resources | `map(string)` | `{}` | no |
 
 ## Outputs
@@ -182,6 +187,7 @@ No modules.
 | <a name="output_instance_profile_id"></a> [instance\_profile\_id](#output\_instance\_profile\_id) | Instance profile's ID |
 | <a name="output_instance_profile_name"></a> [instance\_profile\_name](#output\_instance\_profile\_name) | Name of the instance profile |
 | <a name="output_instance_profile_unique"></a> [instance\_profile\_unique](#output\_instance\_profile\_unique) | Stable and unique string identifying the IAM instance profile |
+| <a name="output_namespace"></a> [namespace](#output\_namespace) | Namespace associated with the Karpenter Pod Identity |
 | <a name="output_node_access_entry_arn"></a> [node\_access\_entry\_arn](#output\_node\_access\_entry\_arn) | Amazon Resource Name (ARN) of the node Access Entry |
 | <a name="output_node_iam_role_arn"></a> [node\_iam\_role\_arn](#output\_node\_iam\_role\_arn) | The Amazon Resource Name (ARN) specifying the node IAM role |
 | <a name="output_node_iam_role_name"></a> [node\_iam\_role\_name](#output\_node\_iam\_role\_name) | The name of the node IAM role |
@@ -189,4 +195,5 @@ No modules.
 | <a name="output_queue_arn"></a> [queue\_arn](#output\_queue\_arn) | The ARN of the SQS queue |
 | <a name="output_queue_name"></a> [queue\_name](#output\_queue\_name) | The name of the created Amazon SQS queue |
 | <a name="output_queue_url"></a> [queue\_url](#output\_queue\_url) | The URL for the created Amazon SQS queue |
+| <a name="output_service_account"></a> [service\_account](#output\_service\_account) | Service Account associated with the Karpenter Pod Identity |
 <!-- END OF PRE-COMMIT-TERRAFORM DOCS HOOK -->

modules/karpenter/main.tf

@@ -411,6 +411,21 @@ resource "aws_iam_role_policy_attachment" "controller_additional" {
   policy_arn = each.value
 }
 
+################################################################################
+# Pod Identity Association
+################################################################################
+
+resource "aws_eks_pod_identity_association" "karpenter" {
+  count = local.create_iam_role && var.enable_pod_identity && var.create_pod_identity_association ? 1 : 0
+
+  cluster_name    = var.cluster_name
+  namespace       = var.namespace
+  service_account = var.service_account
+  role_arn        = aws_iam_role.controller[0].arn
+
+  tags = var.tags
+}
+
 ################################################################################
 # Node Termination Queue
 ################################################################################

modules/karpenter/outputs.tf

@@ -96,3 +96,17 @@ output "instance_profile_unique" {
   description = "Stable and unique string identifying the IAM instance profile"
   value       = try(aws_iam_instance_profile.this[0].unique_id, null)
 }
+
+################################################################################
+# Pod Identity
+################################################################################
+
+output "namespace" {
+  description = "Namespace associated with the Karpenter Pod Identity"
+  value       = var.namespace
+}
+
+output "service_account" {
+  description = "Service Account associated with the Karpenter Pod Identity"
+  value       = var.service_account
+}

modules/karpenter/variables.tf

@@ -138,6 +138,28 @@ variable "irsa_assume_role_condition_test" {
   default     = "StringEquals"
 }
 
+################################################################################
+# Pod Identity Association
+################################################################################
+# TODO - Change default to `true` at next breaking change
+variable "create_pod_identity_association" {
+  description = "Determines whether to create pod identity association"
+  type        = bool
+  default     = false
+}
+
+variable "namespace" {
+  description = "Namespace to associate with the Karpenter Pod Identity"
+  type        = string
+  default     = "kube-system"
+}
+
+variable "service_account" {
+  description = "Service account to associate with the Karpenter Pod Identity"
+  type        = string
+  default     = "karpenter"
+}
+
 ################################################################################
 # Node Termination Queue
 ################################################################################