Event sourcing makes errors worse
Mark Green


How do you handle bugs on the command side. Say, there’s a missing field in some of the events and we cannot replay all of the events in order to rehydrate the aggregate, which means all of the next commands are going to be rejected (by having this exception from the replay logic). How do you handle reconciliation of the already rejected commands? Persistent queues for the commands and enabling a retry logic until we fix the events replay logic?