Event Routing offers a reliability setting that can be defined per service. When you set this option to true for your service, Event Routing guarantees the delivery of each event to the service. Event Routing operates in an asynchronous mode, which means that when it sends an event, it does not wait for an immediate acknowledgment of successful delivery before sending the next event. As a result, if the reliability parameter for a service is set to true, and the runtime where Event Routing is embedded crashes, any events that have been sent before the crash but have not been acknowledged will be redelivered after the restart of the runtime.

If the reliability option is set to false, the delivery of events is not guaranteed. In this case, if the runtime where Event Routing is embedded crashes, no events will be redelivered. Also, if the maximum storage capacity of the event channel is reached, the oldest undelivered events will be discarded in favor of the new events that arrive.

For more information about configuring default maximum storage capacity, see Configuring Storage Capacity for Event Routing. For more information about setting up storage capacity per event type, see Configuring Event Type Associations.