the e2e-aws-ovn-windows job is nearly perma-fail and it's a required job.

it needs to be updated to improve the pass rate or just removed if nobody cares
about it.

the issues are mostly unknown, but the few failures I dug in to (example) are
due to some windows pod not able to get a response to a curl request. The problem
could be anywhere, the test, ovnk networking, cloud infra, etc.

slack thread started here asking for pointers.