test: larger eks image (cloudnative-pg#7053)

Drain E2E on EKS can fail because the operator can become unschedulable
during a drain due to cpu requirements not available. Use a larger
machine.

Closes cloudnative-pg#7052

Signed-off-by: Francesco Canovai <francesco.canovai@enterprisedb.com>

Author: fcanovai

hack/e2e/eks-cluster.yaml.template

@@ -11,7 +11,7 @@ iam:
 
 managedNodeGroups:
   - name: default
-    instanceType: m5.large
+    instanceType: c6g.xlarge
     desiredCapacity: 3
 
 addons: