Transactions are an easy technique for software developers to achieve data consistency. Changes are all committed or rollback, so your data is always aligned and other sessions see no incomplete information. But in distributed microservices environments, this cannot be achieved because data can be spread over several data sources and types of data sources, or we cannot scale distributed transactions very well.
To eventually achieve the data consistency, the SAGA pattern is very popular and describes a procedure to achieve it using the concept of a compensatable action. The session goes a bit deeper into these problems and concepts and shows you with several examples how you can achieve eventual consistency within microservices with the use of the MicroProfile Long Running Action specification which is developed to address this.
Video producer: https://voxxeddays.com/