hi I'm new in Flink development :) I want to build a graph stream that do the following steps: 1. process single events 2. At some point, Group and aggregated the Events to different groups, once done - process in new graph 3. At some point, Group all groups to large group - once done - processing in new graph. Is Flink is good choice in that kind of processing? And if so - should I create a master app that do the aggregation and triggers a new job each time events done to aggregate, or better use Flink for this aggregation parts?