Network does not exactly seem to be interrupted, but It does not make me feel comfortable introducing this into production clusters. There is really not anything i can see in the logs or debug logs that exactly points out the issue, but i can retrieve those. I have tested: -Applying resource reserve/limits -Ensured no local firewalls on the nodes running -Tested updating MTU settings on ubuntu and calico-node config to match the MTU value of the switches (9000/8950) -increasing timeout to 10 and 15 seconds for the readiness check