Stream Management API may report incomplete number of nodes in the cluster

All nodes are reporting their stream configurations and metrics, data is aggregated by a computation that is exposing the result in the KV store, stream and scale endpoints are using this information.

The aggregation is done by IP which is not good in the case nodes are deployed inside docker compose where the local IP can be duplicated on different hosts, this results in an invalid number of nodes reported by endpoints.

Data must be aggregated by nodeId which must be unique in the cluster instead of their IP.