A reliable way to ensure data consistency across systems
Thanks for the nice write-up.
I was wondering what if a failure occurs while consuming the event from the queue / processing the event?
Then we'll have to undo the changes in the service that created this event right? Should that be also handled in a similar manner?
It should depend on the failure type and reason. Our domain has successfully pushed all the data successfully. Consumption will not fall in the publishers domain, so the consumer should decide how to handle the error and not the producer.
Thanks for the nice write-up.
I was wondering what if a failure occurs while consuming the event from the queue / processing the event?
Then we'll have to undo the changes in the service that created this event right? Should that be also handled in a similar manner?
It should depend on the failure type and reason. Our domain has successfully pushed all the data successfully. Consumption will not fall in the publishers domain, so the consumer should decide how to handle the error and not the producer.