[SDC distribution broken]

History: DCAEGEN2 has a service (service-change-handler) that connects to DMAAP, but gets its DMAAP resolution through a request to SDC. (getUeb...), hence relies on SDC configuration to get the DMAAP IPs. Before DCAEGEN2 was introduced, SDC was configured with Kubernetes specific service resolution for DMAAP (dmaap.onap-message-router).
With the introduction of DCAEGEN2, as the service-change-handler running in a VM outside the Kubernetes cluster, the resolution couldn't work with the K8S specific one, hence had to be change to using the k8s external ip. See (https://gerrit.onap.org/r/#/c/26647/5/kubernetes/config/docker/init/src/config/sdc/environments/AUTO.json and https://gerrit.onap.org/r/#/c/26647/5/kubernetes/config/docker/init/config-init.sh l 223

As DCAEGEN2 requires to talk to some OOM services using the service port (hardcoded), and not the NodePort (k8s introduced notion), the design to integrate DCAEGEN2 in OOM was using a reverse proxy to lookup the service name for a given host, on a given port.

The thing is, OOM, kubernetes based, can have services, application being restarted hence their resolution must happen each time a request is made, and should not be cashed.

This fix will update the nginx configuration so it can retrieve the service ip dynamically.