Having just completed a project using Kafka as an event pipeline and a data store, one of the issues we found was that consumer polling takes a large chunk of r

Having just completed a project using Kafka as an event pipeline and a data store, one of the issues we found was that consumer polling takes a large chunk of resources. Having a push model for consumption would certainly remove some of the complexity we had to deal with for scaling out consumption.