This is a pretty big PR, the goal is to provide better support for error scenarios, shutdown behaviour and overall consumer configurations. In the previous versions we were relying on kafka's auto-commit policy, which is okay, but depending on the use case we may want to have tighter control over how we commit the processed offsets.
Also this version introduces some "breaking" changes on the publicly exposed api for the service, namely when getting the consumer and producer instances: The idea is that the producer and consumer each have their own properties to enable finer grained controls
This is a pretty big PR, the goal is to provide better support for error scenarios, shutdown behaviour and overall consumer configurations. In the previous versions we were relying on kafka's auto-commit policy, which is okay, but depending on the use case we may want to have tighter control over how we commit the processed offsets.
Also this version introduces some "breaking" changes on the publicly exposed api for the service, namely when getting the consumer and producer instances: The idea is that the producer and consumer each have their own properties to enable finer grained controls