feat: Add resizePolicy to chart (#60)

* feat: Add `resizePolicy` to chart for in-place pod vertical scaling

Signed-off-by: younsl <cysl@kakao.com>

* docs: Add resizePolicy to README configuration table

Signed-off-by: younsl <cysl@kakao.com>

* fix: align README parameter order with helm-docs generated output

The `resizePolicy` parameter was listed after `resources` in the README
configuration table for both `dcgmAgent` and `nodeAgent` sections.
However, helm-docs generates the table in alphabetical order, placing
`resizePolicy` before `resources`. This mismatch caused the `make generate`
CI check to fail because the generated output did not match the committed
README.

Reorder the parameter rows to match helm-docs alphabetical sorting:
- dcgmAgent.resizePolicy before dcgmAgent.resources
- nodeAgent.resizePolicy before nodeAgent.resources

Signed-off-by: younsl <cysl@kakao.com>

---------

Signed-off-by: younsl <cysl@kakao.com>

Author: younsl

charts/configuration.schema.json

@@ -45,6 +45,9 @@
                 "resources": {
                     "$ref": "#/definitions/Resources"
                 },
+                "resizePolicy": {
+                    "$ref": "#/definitions/ResizePolicy"
+                },
                 "tolerations": {
                     "$ref": "#/definitions/Tolerations"
                 },
@@ -93,6 +96,9 @@
                 "resources": {
                     "$ref": "#/definitions/Resources"
                 },
+                "resizePolicy": {
+                    "$ref": "#/definitions/ResizePolicy"
+                },
                 "tolerations": {
                     "$ref": "#/definitions/Tolerations"
                 }
@@ -141,6 +147,27 @@
                 }
             }
         },
+        "ResizePolicy": {
+            "title": "ResizePolicy",
+            "type": "array",
+            "description": "Container resize policy for in-place pod vertical scaling (requires Kubernetes 1.33+)",
+            "default": [],
+            "items": {
+                "type": "object",
+                "additionalProperties": false,
+                "properties": {
+                    "resourceName": {
+                        "type": "string",
+                        "enum": ["cpu", "memory"]
+                    },
+                    "restartPolicy": {
+                        "type": "string",
+                        "enum": ["NotRequired", "RestartContainer"]
+                    }
+                },
+                "required": ["resourceName", "restartPolicy"]
+            }
+        },
         "Tolerations": {
             "title": "Tolerations",
             "type": "array",

charts/eks-node-monitoring-agent/README.md

@@ -46,6 +46,7 @@ The following table lists the configurable parameters for this chart and their d
 | dcgmAgent.image.tag | string | `"4.5.2-4.8.1-ubuntu22.04"` | Image tag for the dcgm-exporter |
 | dcgmAgent.podAnnotations | object | `{}` | Pod annotations applied to the dcgm exporter |
 | dcgmAgent.podLabels | object | `{}` | Pod labels applied to the dcgm exporter |
+| dcgmAgent.resizePolicy | list | `[]` | Container resize policy for in-place pod vertical scaling (requires Kubernetes 1.33+) |
 | dcgmAgent.resources | object | `{}` | Container resources for the dcgm deployment |
 | dcgmAgent.tolerations | list | `[]` | Deployment tolerations for the dcgm |
 | fullnameOverride | string | `"eks-node-monitoring-agent"` | A fullname override for the chart |
@@ -66,6 +67,7 @@ The following table lists the configurable parameters for this chart and their d
 | nodeAgent.monitors | object | `{}` | Per-monitor configuration keyed by plugin name. See the main README for details. |
 | nodeAgent.podAnnotations | object | `{}` | Pod annotations applied to the eks-node-monitoring-agent |
 | nodeAgent.podLabels | object | `{}` | Pod labels applied to the eks-node-monitoring-agent |
+| nodeAgent.resizePolicy | list | `[]` | Container resize policy for in-place pod vertical scaling (requires Kubernetes 1.33+) |
 | nodeAgent.resources | object | `{"limits":{"cpu":"250m","memory":"200Mi"},"requests":{"cpu":"10m","memory":"30Mi"}}` | Container resources for the eks-node-monitoring-agent |
 | nodeAgent.securityContext | object | `{"capabilities":{"add":["NET_ADMIN"]},"privileged":true}` | Container Security context for the eks-node-monitoring-agent |
 | nodeAgent.tolerations | list | `[{"operator":"Exists"}]` | Deployment tolerations for the eks-node-monitoring-agent |

charts/eks-node-monitoring-agent/templates/daemonset.yaml

@@ -65,6 +65,10 @@ spec:
             httpGet:
               path: /healthz
               port: 8002
+          {{- with .Values.nodeAgent.resizePolicy }}
+          resizePolicy:
+            {{- toYaml . | nindent 12 }}
+          {{- end }}
           {{- with .Values.nodeAgent.resources }}
           resources:
             {{- toYaml . | nindent 12 }}

charts/eks-node-monitoring-agent/templates/dcgm-daemonset.yaml

@@ -46,6 +46,10 @@ spec:
           imagePullPolicy: {{ .Values.dcgmAgent.image.pullPolicy }}
           command: ["/bin/sh"]
           args: ["-c", "nv-hostengine -n -b ALL || true; sleep infinity"]
+          {{- with .Values.dcgmAgent.resizePolicy }}
+          resizePolicy:
+            {{- toYaml . | nindent 12 }}
+          {{- end }}
           {{- with .Values.dcgmAgent.resources }}
           resources:
             {{- toYaml . | nindent 12 }}

charts/eks-node-monitoring-agent/values.yaml

@@ -76,6 +76,8 @@ nodeAgent:
     limits:
       cpu: 250m
       memory: 200Mi
+  # -- Container resize policy for in-place pod vertical scaling (requires Kubernetes 1.33+)
+  resizePolicy: []
   # -- Container Security context for the eks-node-monitoring-agent
   securityContext:
     privileged: true
@@ -138,6 +140,8 @@ dcgmAgent:
               - amd64
   # -- Container resources for the dcgm deployment
   resources: {}
+  # -- Container resize policy for in-place pod vertical scaling (requires Kubernetes 1.33+)
+  resizePolicy: []
   # -- Deployment tolerations for the dcgm
   tolerations: []
   # -- Pod labels applied to the dcgm exporter