Description of problem:
 * Customer has deployed an AWS 4.12.33 cluster and is observing that networkpolicy creation for target namespaces are not working as expected.
 * A deny by default netpol does not prevent traffic from being routed to backends via router-default pods (route).
 *  

Version-Release number of selected component (if applicable):
 * AWS
 * 4.12.33

How reproducible:
 * every time on this cluster and one other
 *  

Steps to Reproduce:

1. Create new namespace

2. deploy an application that serves a 200 response for looped curl handling

3. loop a curl to the route after exposing it for access testing:

 while true; do curl --noproxy '*' -k -w "dnslookup: %\{time_namelookup} | connect: %\{time_connect} | appconnect: %\{time_appconnect} | pretransfer: %\{time_pretransfer} | starttransfer: %\{time_starttransfer} | total: %\{time_total} | size: %\{size_download} | response: %\{response_code}\n" -o /dev/null -s [https://URL;|https://url;/] sleep 1; done

4. Observe 200 responses from the route

5. Create a deny-by-default networkpolicy and apply it to the namespace:

kind: NetworkPolicy
apiVersion: networking.k8s.io/v1
metadata:
  name: deny-by-default
  namespace: <test-namespace>
spec:
  podSelector: {}
  ingress: []{code}
6. netpol is applied successfully - traffic is not interrupted to route via looped curl (200's still available).

replicated on local test cluster internally - netpol above immediately stops the 200 response and starts to return a 503 (expected) on internal test platform.

Actual results:
 * traffic is not obstructed/modified by netpols when routing via route
 *  

Expected results:
 * traffic should be dropped unless explicitly set by netpol if even one netpol is applied successfully to a namespace
 *  

Additional info:

Please fill in the following template while reporting a bug and provide as much relevant information as possible. Doing so will give us the best chance to find a prompt resolution.

Affected Platforms:
 * AWS (customer env observed- cannot replicate with resourcehub/quicklab)

Data available in first comment with specific findings + logs