Describe the bug
Currently when an error is returned in saga event handling, the event is not retried. Only if the application is restarted the event is retried.
In that case when an event caused an error, and subsequently another event is handled successfully, the previous, unsuccessful event will never be retried, not even after restart.
To Reproduce
- Create an kafka eventbus
- Create a saga
- Let the saga handle an event and return an error
- Notice, error will be logged
- Event is never retried
Expected behavior
- Create an kafka eventbus
- Create a saga
- Let the saga handle an event and return an error
- Notice, error will be logged
- Event is retried
- Until event is successful do not proceed to next event.
Further
Currently it seems implementation specific what the semantics of the eventbus are with regards to in-orderness, retrying, lossyness.
The Kafka eventbus seems to be in-order, but lossy.
Describe the bug
Currently when an error is returned in saga event handling, the event is not retried. Only if the application is restarted the event is retried.
In that case when an event caused an error, and subsequently another event is handled successfully, the previous, unsuccessful event will never be retried, not even after restart.
To Reproduce
Expected behavior
Further
Currently it seems implementation specific what the semantics of the eventbus are with regards to in-orderness, retrying, lossyness.
The Kafka eventbus seems to be in-order, but lossy.