Hi
i have continuous log messages sent to kafka , i am using the DFP example and transforming the kafka messages periodically into json files and then executing the DFP inference pipeline on the json files i have created.
my question: Is it possible to inference instantly each message and if so what the best practice