Question 1: Does kafka streams processor api allows both single and batch record processing ( or it is always batch record processing internally ) ?
Question 2: What config or setting is needed to set single or batch processing and how to implement batch processing ( snippet, link) using kafka streams processor api ?
Kafka Streams processor api processes a single record at a time initially, but you could use state stores with Transformer for stateful operations for achieving batching.
I guess you need batching by both size and time (e.g. if during 5 seconds batch size not reached, propagate data what we already collected). If so, please take a look at example How to process a KStream in a batch of max size or fallback to a time window. If you need time-based aggregation (and not by size), please take a look at windowing.
Actually, you don't need some specific properties for using processor api (only regular props for Kafka Streams like bootstrap.servers, application.id, auto.offset.reset etc.), and only for batching you need to declare a state store.
