Could you elaborate a little on the traceability part, please? If the different steps of a processing chain are distributed via queues, you rather have an overhead in collecting the information in a central place, I would think.
That is certainly true, but message queues are pretty well suited to tagging and audit logs. So the systems involved tag the messages and an audit repository can accept logs from all services.
I guess this isn't in contrast to a monolith which is easier to log, but in contrast to non-message based microservices I think it's easier to implement the audit logging.
I guess this isn't in contrast to a monolith which is easier to log, but in contrast to non-message based microservices I think it's easier to implement the audit logging.