Ingresses in the "cluster-wide" example test specify the admin Service.

This is in anticipation of removing the rule that a viking service
implements an Ingress if it's the only one in the entire cluster.
This can cause unexpected changes if something unrelated in the
cluster is changed, for example by independent teams working in
different namespaces.

This makes it unnecessary to undeploy the viking service that always
runs in namespace default during e2e tests (which was just such a
dependency).

The docs are not currently up to date on this point.

examples/architectures/clusterwide/coffee-ingress.yaml

@@ -5,6 +5,7 @@ metadata:
   namespace: coffee
   annotations:
     kubernetes.io/ingress.class: "varnish"
+    ingress.varnish-cache.org/varnish-svc: "kube-system/varnish-ingress-admin"
 spec:
   rules:
   - host: coffee.example.com

examples/architectures/clusterwide/other-ingress.yaml

@@ -5,6 +5,7 @@ metadata:
   namespace: other
   annotations:
     kubernetes.io/ingress.class: "varnish"
+    ingress.varnish-cache.org/varnish-svc: "kube-system/varnish-ingress-admin"
 spec:
   backend:
     serviceName: other-svc

examples/architectures/clusterwide/tea-ingress.yaml

@@ -5,6 +5,7 @@ metadata:
   namespace: tea
   annotations:
     kubernetes.io/ingress.class: "varnish"
+    ingress.varnish-cache.org/varnish-svc: "kube-system/varnish-ingress-admin"
 spec:
   rules:
   - host: tea.example.com