Adjust resource requests and autoscaling of calico-node side car

containers.

The side car containers use negligible amount of resources compared to
`calico-node`. Therefore, requests likely just waste resources.
Similarly, they fall below the minimum threshold of VPA. Hence,
autoscaling them does not make sense.
`minAllowed` and `maxAllowed` of `calico-node` also are not that useful.
Therefore, they are removed as well.

charts/internal/calico/templates/node/daemonset-calico-node.yaml

@@ -171,10 +171,6 @@ spec:
           env:
           - name: POD_CIDR
             value: {{ .Values.global.podCIDR }}
-          resources:
-            requests:
-              cpu: 10m
-              memory: 50Mi
           command:
           - /bin/sh
           - -c
@@ -238,10 +234,6 @@ spec:
               valueFrom:
                 fieldRef:
                   fieldPath: spec.nodeName
-          resources:
-            requests:
-              cpu: 1m
-              memory: 20Mi
           command:
             - /bin/sh
             - -c

charts/internal/calico/templates/node/vpa.yaml

@@ -14,25 +14,5 @@ spec:
   resourcePolicy:
     containerPolicies:
     - containerName: calico-node
-      minAllowed:
-        memory: 50Mi
-      maxAllowed:
-        memory: 2800Mi
-      controlledValues: RequestsOnly
-    - containerName: "add-snat-rule-to-upstream-dns"
-      minAllowed:
-        cpu: 10m
-        memory: 50Mi
-      maxAllowed:
-        cpu: 1
-        memory: 1G
-      controlledValues: RequestsOnly
-    - containerName: "network-unavailable-condition-ensurer"
-      minAllowed:
-        cpu: 1m
-        memory: 20Mi
-      maxAllowed:
-        cpu: 10m
-        memory: 100Mi
       controlledValues: RequestsOnly
 {{- end }}