ClusterController

[parhamdoustdar]

When we're ready to introduce capacity based scheduling, we should create a Cluster Controller.

Its primary task is to collect capacity information from clusters (by consuming their Node objects) and fill the Cluster.Status with aggregated capacity data for consumption by the Schedule controller.

It could also use the information from the clients to set conditions on the Cluster object to help provide more information to operators. For example, if Shipper has problems connecting to a cluster, this would be the place to surface that information.