je-ik commented on issue #23379: URL: https://github.com/apache/beam/issues/23379#issuecomment-1263683779
Where exactly is the breakage? If the set of open windows is finite (which it is), then there is finite set of open windows with lowest timestamp. The fact that watermark might represent "wall clock" doesn't mean that it has to advance. That is what I meant with non-negative speed wrt. processing time. The watermark can stay still for any (bounded) amount of processing time, which should solve the batch case. It only cannot travel back in time (excluding failure recovery) - i.e. dTp/dTe >= 0 for all event times Te and processing times Tp. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
