fix(spark): lower memory requests to fix scheduling on resource-constrained nodes

Scheduler uses requests (not limits) for placement. Previous 1Gi requests
per pod caused Insufficient memory even with plenty of physical RAM.
New requests: master 256Mi, workers 256Mi, history-server 128Mi.
Limits unchanged. SPARK_WORKER_MEMORY reduced to 512m to match worker budget.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

spark-cluster/base/history-server.yaml

@@ -31,7 +31,7 @@ spec:
           resources:
             requests:
               cpu: 100m
-              memory: 512Mi
+              memory: 128Mi
             limits:
               cpu: 500m
               memory: 1Gi

spark-cluster/base/master.yaml

@@ -29,8 +29,8 @@ spec:
               containerPort: 8080
           resources:
             requests:
-              cpu: 500m
+              cpu: 200m
-              memory: 1Gi
+              memory: 256Mi
             limits:
               cpu: "1"
               memory: 2Gi

spark-cluster/base/workers.yaml

@@ -27,14 +27,14 @@ spec:
             - name: SPARK_WORKER_CORES
               value: "1"
             - name: SPARK_WORKER_MEMORY
-              value: "1g"
+              value: "512m"
           ports:
             - name: webui
               containerPort: 8081
           resources:
             requests:
-              cpu: 500m
+              cpu: 200m
-              memory: 1Gi
+              memory: 256Mi
             limits:
               cpu: "1"
               memory: 2Gi