1. Proposed title of this feature request

OpenShift 4: Unexpected Node IP change alert system

2. What is the nature and description of the request?

Recently a cluster was heavily impacted during the loss of an upstream DHCP server that caused all hosts to lose their DHCP leases for a period of time before the server was repaired. After the DHCP server was brought up, nodes were allocated IPs but not all nodes received the same IP lease they had previously, causing confusion in the cluster. Node reboots did not alleviate, but OVNKUBE DB rebuilds allowed the worker nodes to re-integrate with the cluster with their new addressing after the local network tables were flushed. 

3. Why does the customer need this? (List the business requirements here)

This RFE seeks to reduce downtime and improve response time for customers when a similar scenario occurs. If the Node IP changes from what the node object has listed, the Network CO (or the kubelet on the node itself) should be able to report that information to the customer to take steps to alleviate/rectify the problem: `nodeIP x.x.x.x is different from expected value: y.y.y.y`. This data would improve response time during an outage and reduce impacts in a networking loss scenario. A recent outage was potentially extended by not seeing the nodes IPs had changed for some time after the dhcp upstream issue was resolved. 

4. List any affected packages or components.

OVNkubernetes (ovnkube-node)

Network Operator

Kubelet