fix(homeassistant): reduce memory requests to 256Mi, add node affinity for raspberrypi

Node roger-nucbox-evo-x2 is at 3.78% memory free causing home-assistant
pods to flap Pending/Running. Reduce resource requests to fit within
cluster capacity and prefer scheduling on raspberrypi (63% mem free).

homeassistant/homeassistant.yaml

@@ -55,12 +55,12 @@ spec:
     spec:
       containers:
       - name: home-assistant
-        image: ghcr.io/home-assistant/home-assistant:stable
+        image: ghcr.io/home-assistant/home-assistant:2026.7.4
         resources:
           requests:
-            memory: "512Mi"
+            memory: "256Mi"
           limits:
-            memory: "1Gi"
+            memory: "512Mi"
         ports:
         - containerPort: 8123
         volumeMounts:
@@ -101,6 +101,16 @@ spec:
         - name: configuration
           mountPath: /config-cm
           readOnly: true
+      affinity:
+        nodeAffinity:
+          preferredDuringSchedulingIgnoredDuringExecution:
+          - weight: 100
+            preference:
+              matchExpressions:
+              - key: kubernetes.io/hostname
+                operator: In
+                values:
+                - raspberrypi
       hostNetwork: true
       volumes:
         - name: config