I want to make an Automation where the flow goes first trough an workspace which reads data from an database. Depending on the feature type it then goes to different Automation Writers. Each of these ports are then connected to its own workspace so it can be run parallel. Eventually they merge back to another workspace which writes back the data to the database.
The problem I have now is that thousands of features reach the Automation Writer thus kicking of the next workspace thousands of time. Obviously that is not what I want. What is best practice to collect and group these features and pass them on as one feature.
I tried using the Aggregator Transformer and adding the parameters to a list, but it seems that it is not supported by GeoJSON.
I use FME Server 2020.2