Hi Salva I think this depends on what the relationship between you output and input events. If the output ones are just simple wrapper of input ones, e.g. adding some simple properties or just read from one place and write to another place, I think the output events could hold time which is inherited from input ones. That is to say, event-time semantics might be more proper. On the other hand, if the output events have more independent relationship with input ones, and those tasks in Flink TM could be treated as the event generator, I think you can make the time as the processing time when generating them. I think there is no absolute rules and all depends on your actual scenarios.
Best Yun Tang ________________________________ From: Salva Alcántara <salcantara...@gmail.com> Sent: Monday, April 20, 2020 2:03 To: user@flink.apache.org <user@flink.apache.org> Subject: Modelling time for complex events generated out of simple ones My flink application generates output (complex) events based on the processing of (simple) input events. The generated output events are to be consumed by other external services. My application works using event-time semantics, so I am bit in doubt regarding what should I use as the output events' timestamp. Should I use: - the processing time at the moment of generating them? - the event time (given by the watermark value)? - both? For my use case, I am using both for now. But maybe you can come up with examples/justifications for each of the given options. -- Sent from: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/
