feat(rolling-update): add hyperpod-patching subchart with RBAC for pod eviction (#91)

Co-authored-by: Vinay Devadiga <[email protected]>

Author: divincode

helm_chart/HyperPodHelmChart/Chart.yaml

@@ -74,4 +74,8 @@ dependencies:
   - name: team-role-and-bindings
     version: "0.1.0"
     repository: "file://charts/team-role-and-bindings"
-    condition: team-role-and-bindings.enabled
+    condition: team-role-and-bindings.enabled
+  - name: hyperpod-patching
+    version: "0.1.0"
+    repository: "file://charts/hyperpod-patching"
+    condition: hyperpod-patching.enabled

helm_chart/HyperPodHelmChart/charts/hyperpod-patching/Chart.yaml

@@ -0,0 +1,5 @@
+apiVersion: v2
+name: hyperpod-patching
+description: A subchart for RBAC used by HyperPod patching workflows
+version: 0.1.0
+appVersion: "1.0"

helm_chart/HyperPodHelmChart/charts/hyperpod-patching/templates/clusterrole.yaml

@@ -0,0 +1,11 @@
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: hyperpod-patching
+rules:
+  - apiGroups: [""]
+    resources: ["pods"]
+    verbs: ["list"]
+  - apiGroups: [""]
+    resources: ["pods/eviction"]
+    verbs: ["create"]

helm_chart/HyperPodHelmChart/charts/hyperpod-patching/templates/clusterrolebinding.yaml

@@ -0,0 +1,12 @@
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: hyperpod-patching
+subjects:
+  - kind: User
+    name: hyperpod-service-linked-role
+    apiGroup: rbac.authorization.k8s.io
+roleRef:
+  kind: ClusterRole
+  name: hyperpod-patching
+  apiGroup: rbac.authorization.k8s.io

helm_chart/HyperPodHelmChart/charts/hyperpod-patching/values.yaml

@@ -259,4 +259,6 @@ health-monitoring-agent:
 deep-health-check:
   enabled: true
 job-auto-restart:
-  enabled: true   
+  enabled: true
+hyperpod-patching:
+  enabled: true

helm_chart/HyperPodHelmChart/values.yaml

@@ -26,6 +26,7 @@ chmod 700 get_helm.sh
 | neuron-device-plugin         | Deploys the AWS Neuron device plugin for Kubernetes, enabling support for AWS Inferentia chips to accelerate machine learning model inference workloads.                                | Yes               |
 | storage                      | Manages persistent storage resources for Kubernetes applications, ensuring that data is retained and accessible across pod restarts and cluster upgrades.                               | No                |
 | training-operators           | Installs operators for managing various machine learning training jobs, such as TensorFlow, PyTorch, and MXNet, providing native Kubernetes support for distributed training workloads. | Yes               |
+| HyperPod patching            | Deploys the RBAC and controller resources needed for orchestrating rolling updates and patching workflows in SageMaker HyperPod clusters. Includes pod eviction and node monitoring.    | Yes               |
 
 ## 3. Test the Chart Locally