In the Console's dashboards page, when opening either "Kubernetes / Networking / Pod" dashboard or "Kubernetes / Networking / Namespace (Pod)", the only available namespace is "openshift-cluster-samples-operator". Other namespaces are not listed there.

Inspecting the network metrics such as container_network_receive_bytes_total (but all cadvisor network metrics seem affected), for some reason, only the one for openshift-cluster-samples-operator seems correctly labelled with pod & namespace. All the other ones miss these labels.

It seems like the instance ip of these metrics refers to the node IP and not the pod IP, which would explain why they aren't labelled with the pod info. Cf attached screenshots 

ocp 4.13-ec3

Always

1. As an admin, open the openshift console, make sure you are in the Admin perspective
2. navigate to "Observe" > "Dashboards"
3. select dashboard "Kubernetes / Networking / Pod" 
4. Check namespaces available from the dropdown

Only openshift-cluster-samples-operator is available

All namespaces generating pods traffic should be listed

I did not test on the latest build (ocp-4.13 ec4)
I'm using OVN-Kubernetes on my cluster, on AWS. I don't have any specific setup beside that - all defaults settings.

It might help to pay attention to the interface labels: they might indicate whether the metric relates to pod-overlay network or host network. E.g. there are interfaces like "br-ex" => external / "br-int" => internal, some randomish characters like "57f21aaebc94371" are pods interfaces. Interestingly, the only workload that is correctly labelled, openshift-cluster-samples-operator, uses eth0.