When I create a stream composed of multiple modules, are there any guarantees provided on the event delivery to each module. Is there a possibility of events being replayed?
What if the processor module takes a long time to process one event but crashes in the middle of the processing. What happens to that event when the module is restarted in a container?
See the documentation re: Message Bus Configuration and specifically
Error Handling (Message Delivery Failures)
.The retry configuration can be modified at the bus or individual module/stream level.