From 541565de483e58a365a50f99c39f3520b06acf22 Mon Sep 17 00:00:00 2001
From: OG T <ogt@WOOOMacMiniM4.local>
Date: Sat, 28 Mar 2026 22:03:54 +0800
Subject: [PATCH] feat(k4.2): Descheduler for pod rebalancing

- Deploy Descheduler v0.30.1 as CronJob
- Schedule: Every 2 hours
- Policies enabled:
  - LowNodeUtilization: rebalance when node < 20% usage
  - RemoveDuplicates: spread replicas across nodes
  - RemovePodsViolatingNodeAffinity: enforce affinity rules

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
---
 k8s/descheduler/descheduler.yaml | 135 +++++++++++++++++++++++++++++++
 1 file changed, 135 insertions(+)
 create mode 100644 k8s/descheduler/descheduler.yaml

diff --git a/k8s/descheduler/descheduler.yaml b/k8s/descheduler/descheduler.yaml
new file mode 100644
index 000000000..61b6485bd
--- /dev/null
+++ b/k8s/descheduler/descheduler.yaml
@@ -0,0 +1,135 @@
+# =============================================================================
+# Descheduler
+# =============================================================================
+# K4.2 2026-03-28: Pod rebalancing for load distribution
+# Deployed by: Claude Code (首席架構師)
+# Schedule: Every 2 hours (cron)
+# =============================================================================
+---
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: descheduler
+  labels:
+    app.kubernetes.io/name: descheduler
+---
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: descheduler
+  namespace: descheduler
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: descheduler
+rules:
+  - apiGroups: [""]
+    resources: ["nodes"]
+    verbs: ["get", "watch", "list"]
+  - apiGroups: [""]
+    resources: ["pods"]
+    verbs: ["get", "watch", "list", "delete"]
+  - apiGroups: [""]
+    resources: ["pods/eviction"]
+    verbs: ["create"]
+  - apiGroups: ["scheduling.k8s.io"]
+    resources: ["priorityclasses"]
+    verbs: ["get", "watch", "list"]
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: descheduler
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: descheduler
+subjects:
+  - kind: ServiceAccount
+    name: descheduler
+    namespace: descheduler
+---
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: descheduler-policy
+  namespace: descheduler
+data:
+  policy.yaml: |
+    apiVersion: descheduler/v1alpha2
+    kind: DeschedulerPolicy
+    profiles:
+      - name: default
+        pluginConfig:
+          - name: LowNodeUtilization
+            args:
+              targetThresholds:
+                cpu: 50
+                memory: 50
+                pods: 50
+              thresholds:
+                cpu: 20
+                memory: 20
+                pods: 20
+          - name: RemovePodsViolatingNodeAffinity
+            args:
+              nodeAffinityType:
+                - requiredDuringSchedulingIgnoredDuringExecution
+          - name: RemoveDuplicates
+            args:
+              excludeOwnerKinds:
+                - DaemonSet
+        plugins:
+          balance:
+            enabled:
+              - LowNodeUtilization
+              - RemoveDuplicates
+          deschedule:
+            enabled:
+              - RemovePodsViolatingNodeAffinity
+---
+apiVersion: batch/v1
+kind: CronJob
+metadata:
+  name: descheduler
+  namespace: descheduler
+  labels:
+    app.kubernetes.io/name: descheduler
+    app.kubernetes.io/version: "0.30.1"
+spec:
+  schedule: "0 */2 * * *"  # Every 2 hours
+  concurrencyPolicy: Forbid
+  successfulJobsHistoryLimit: 3
+  failedJobsHistoryLimit: 3
+  jobTemplate:
+    spec:
+      template:
+        metadata:
+          labels:
+            app.kubernetes.io/name: descheduler
+        spec:
+          serviceAccountName: descheduler
+          restartPolicy: Never
+          containers:
+            - name: descheduler
+              image: registry.k8s.io/descheduler/descheduler:v0.30.1
+              imagePullPolicy: IfNotPresent
+              command:
+                - /bin/descheduler
+                - --policy-config-file=/policy-dir/policy.yaml
+                - --v=3
+              resources:
+                limits:
+                  cpu: 200m
+                  memory: 128Mi
+                requests:
+                  cpu: 100m
+                  memory: 64Mi
+              volumeMounts:
+                - name: policy-volume
+                  mountPath: /policy-dir
+          volumes:
+            - name: policy-volume
+              configMap:
+                name: descheduler-policy