h-a/{piper,whisper}: Prefer x86_64 nodes

Without a node affinity rule, Kubernetes applies equal weight to the
"big" x86_64 nodes and the "small" aarch64 ones.  Since we would really
rather Piper and Whisper _not_ run on a Raspberry Pi, we need the rule
to express this.

home-assistant/piper.yaml

@@ -36,6 +36,16 @@ spec:
         app.kubernetes.io/name: piper
         app.kubernetes.io/part-of: home-assistant
     spec:
+      affinity:
+        nodeAffinity:
+          preferredDuringSchedulingIgnoredDuringExecution:
+          - weight: 100
+            preference:
+              matchExpressions:
+              - key: kubernetes.io/arch
+                operator: In
+                values:
+                - amd64
       containers:
       - name: piper
         image: docker.io/rhasspy/wyoming-piper:1.3.2

home-assistant/whisper.yaml

@@ -36,6 +36,16 @@ spec:
         app.kubernetes.io/name: whisper
         app.kubernetes.io/part-of: home-assistant
     spec:
+      affinity:
+        nodeAffinity:
+          preferredDuringSchedulingIgnoredDuringExecution:
+          - weight: 100
+            preference:
+              matchExpressions:
+              - key: kubernetes.io/arch
+                operator: In
+                values:
+                - amd64
       containers:
       - name: whisper
         image: docker.io/rhasspy/wyoming-whisper:1.0.0