The primary benefit of the solution is that unrecoverable events, which are usually caused by bugs, will not delay the processing of legitimate events. Stream processors handle many different types of events. Some are of no interest to a specific processor and will be ignored and advance the checkpoint. Other events follow a different processing path that does not encounter an issue, while others follow the same path but the events themselves are valid. This increases stability and resilience because stream processors are able to continue to make forward progress in the face of failures and potentially self-heal.
One drawback is that a large number of fault events can be produced that will require manual intervention and ...
