On Google Cloud Platform, I am trying to have an RTMP stream go to a load balancer, and then have that load balancer distribute the stream to all the instances behind a managed service. Meaning that the LB shouldn't use things like Round Robin or CPU utilization, it should just go to all the instances.
The use case scenario is the infrastructure is designed to be a push (over pull). In my managed instances, I have several Node Media Servers that auto-scale based CPU utilized because the RTMP stream is be converted to HLS. Now each instance will require its own stream to be pushed into it for the conversion to occur and the user to recieve the stream content.
How can this done at GCP if not with a load balancer?