Improve connectivity probe for dual-stack systems #1411
Description
Currently there is one connectivity probe that checks ping towards default gateway. This creates a problem in dual-stack systems because we can only mark probe based on the logical AND or OR for those two checks. Whatever our choice is, we can't easily see when we break only one of the IP stacks.
Ideally we would have ping4 and ping6 probe that tracks IP stacks separately. That way a change to the probe status after applying NNCP would give a clear signal what broke.
A side-effect would be that in single-stack clusters one of the probes always fails but this is acceptable.