Motivation and background:
When we face failures in VLAN-involved tests, we usually cannot tell immediately if it's because of lack of VLAn support in the cluster (for example - a specific VLAN tag, which is assumed to be supported in the cluster, is not really supported). A VLAN sanity test can avoid that.
As part of the network sanity tests performed in the beginning of the tier-2 tests, we might want to add this VLAN sanity check, by verifying in advance that a successful connectivity is possible over these VLAN tags.

Potential challenges:
1
How to recognize the tests that require this check (as we wouldn't want to fail the entire build if VLAN is not required in it)?
Most likely that we will mark these tests.

2.
How to verify it?
Sounds to me that the simplest would be creating VLAN interfaces over the nodes, and verify connectivity between them, but there are few problems with this approach:
a - how to create the VLANs in cluster with no k8s-nmstate?
b - in large-scale clusters, this might take forever.
c - BMs have a range of 20 supported VLAN tags - see the next section.

3.
Should we verify all the VLAN tags that are declared to be available in the cluster?
This might be problematic in most of the bare-metals, where the VLAN range is of 20 tags (1000-1019). This check might take a very long time.
Possible solutions:
a - Assuming that we eventually use a very limited number of the VLAN tags in the test (I think that no more than 3 different tags), we can apply the sanity only of 4-5 tags.
b - Create all the VLAN interfaces, with an IP address assigned to each, in one NNCP. But again - this will not be possible in clusters without k8s-nmstate.