Status

Current state: Under Discussion

Discussion thread: TBD

JIRA:   Unable to render Jira issues macro, execution error. Unable to render Jira issues macro, execution error.

Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).

Motivation

The main motivation of this KIP is stated in the related JIRA: "Today the downstream sub-topology's parallelism (aka the number of tasks) are purely dependent on the upstream sub-topology's parallelism, which ultimately depends on the source topic's num.partitions. However this does not work perfectly with dynamic scaling scenarios". By delegating the stream topology power to create repartition topic with customized number of partitions gives user more flexibility. Also, for API like #to or #through, KStream has no access to sink topic creation, which means user has to create their own connecting Kafka topic for every single new application which is cumbersome. Thus we are proposing extending the capability of #Produced API to automatically create topics when the target topic has not created.

Public interfaces

We shall expand the Produced API to contain topicName and numPartitions: 

public class Produced<K, V> {
	protected Serde<K> keySerde;
	protected Serde<V> valueSerde;
	protected StreamPartitioner<? super K, ? super V> partitioner;
	protected final String topicName; // new
	protected final String numPartitions; // new



	public static <K, V> Produced<K, V> with(final Serde<K> keySerde,
                                         	 final Serde<V> valueSerde,
                                             final StreamPartitioner<? super K, ? super V> partitioner,
											 final topicName,	
										     final Integer numPartitions);
}

We also want to expand Grouped API with a numPartition configuration. 

public class Grouped<K, V> {
 	protected final Serde<K> keySerde;
 	protected final Serde<V> valueSerde;
 	protected final String name;
	protected final String numPartitions; // new

	public static <K, V> Grouped<K, V> with(final String name,
    	                                    final Serde<K> keySerde,
        	                                final Serde<V> valueSerde,
											final Integer numPartitions);
}

Proposed Changes

 When configured with topic name and partition,  KStream application will first issue the topic lookup request and check whether the target topic is already up and running. If the target topic already exists, KStream application will log the error, and still continue the initialization.

For grouping operation,  repartition topic gets a hint on how many partitions the new topic should be created. If repartition topic is not created yet, create one with specified numPartitions, otherwise use the upstream topic partition size. If repartition logic is created, we shall issue a request to request to change the repartition topic size if current topic numPartitions != configured numPartitions. Note that there could be potential data loss, so one open question is whether we should prevent destructive operation like this.

Backward Compatibility 

Rejected Alternatives

N/A