Fork me on GitHub

Hi @wparker, thanks for your explanation. What I had thought was to try to treat a session as a source of streaming data. As soon as a new fact is inserted/retracted another job would examine it and take another necessary actions. I see that I can easily obtain a similar effect composing the insert! with sending the fact to a Kafka cluster (for example) but not sure how to deal with the retracts that happens automatically in a clara session.