Strategy for Resource allocation in data lake - Batch and Streaming Jobs

Question

In our bigdata environment, we use Nifi for both Stream processing/ingestion from Kafka and batch ingestions. Given the finite resources, what is a good strategy to optimiz