If we have two different groups of hosts using the same subnet (for example, with different OpenStack projects), we could end creating edges between nodes that are not connected.
Now that should produce a warning (two listeners for the same endpoint).
Maybe a solution for this cases is to run a skydive instance for each subnet and somewho connect those skydives knowing that they are processing different networks?
If we have two different groups of hosts using the same subnet (for example, with different OpenStack projects), we could end creating edges between nodes that are not connected.
Now that should produce a warning (two listeners for the same endpoint).
Maybe a solution for this cases is to run a skydive instance for each subnet and somewho connect those skydives knowing that they are processing different networks?