Kafka Streams Flashcards

What is a 'tumbling window' in Kafka Streams?

Answer: A fixed-size, non-overlapping time window where each record belongs to exactly one window

Tumbling windows partition time into equal-sized, non-overlapping buckets so each event falls into exactly one window.

What is the purpose of `StreamsConfig.APPLICATION_ID_CONFIG`?

Answer: Uniquely identifies the Kafka Streams application and is used as the consumer group ID and internal topic prefix

The application ID serves as the consumer group ID for all internal consumers and is prepended to internal changelog and repartition topic names.

What does the `filter()` operation do in a Kafka Streams topology?

Answer: Passes only records that satisfy a given predicate, dropping the rest

filter() evaluates a Predicate for each record and only forwards records where the predicate returns true.

What is a 'changelog topic' in Kafka Streams?

Answer: A Kafka topic used to back up local state store contents for fault tolerance

Kafka Streams writes all state store updates to a changelog topic so that state can be restored from Kafka if a task migrates to another instance.

What does `flatMapValues()` do in Kafka Streams?

Answer: Transforms each record's value into zero or more values, emitting one record per resulting value

flatMapValues() applies a function to each record's value that returns an iterable, emitting a separate downstream record for each element.

What is a 'session window' in Kafka Streams?

Answer: A dynamic window that groups records separated by an inactivity gap smaller than a configurable threshold

Session windows group events for a key into a session as long as the gap between consecutive events is less than the inactivity gap timeout.