How to maintain atomicity

[surajpanda1582]

Hi Chris

I am having one questions in mind that why a separate process is require to mine the data from transaction log and publish into kafka system. If the service itself after saving the event in event store and issue a commit request then put the event in kafka system, will it not be a feasible solution. What is the drawback with this approach. Mining the transaction log is little unfamiliar to me, so thought of designing the solution with this approach. Please let me know your suggestion.