thanks @milans100 -- I was hoping to avoid that since the data could be easily trivially computed on the server and I wouldn't have to insert a "smart processor" in the pipeline (in my case, I was hoping to have osquery -> kafka -> druid)