software.amazon.kinesis.connectors.flink

Class FlinkKinesisConsumer<T>

java.lang.Object

org.apache.flink.api.common.functions.AbstractRichFunction

org.apache.flink.streaming.api.functions.source.RichParallelSourceFunction<T>

software.amazon.kinesis.connectors.flink.FlinkKinesisConsumer<T>

Type Parameters:

T - the type of data emitted

All Implemented Interfaces:

Serializable, org.apache.flink.api.common.functions.Function, org.apache.flink.api.common.functions.RichFunction, org.apache.flink.api.java.typeutils.ResultTypeQueryable<T>, org.apache.flink.streaming.api.checkpoint.CheckpointedFunction, org.apache.flink.streaming.api.functions.source.ParallelSourceFunction<T>, org.apache.flink.streaming.api.functions.source.SourceFunction<T>

Direct Known Subclasses:

FlinkDynamoDBStreamsConsumer

@PublicEvolving
public class FlinkKinesisConsumer<T>
extends org.apache.flink.streaming.api.functions.source.RichParallelSourceFunction<T>
implements org.apache.flink.api.java.typeutils.ResultTypeQueryable<T>, org.apache.flink.streaming.api.checkpoint.CheckpointedFunction

The Flink Kinesis Consumer is an exactly-once parallel streaming data source that subscribes to multiple AWS Kinesis streams within the same AWS service region, and can handle resharding of streams. Each subtask of the consumer is responsible for fetching data records from multiple Kinesis shards. The number of shards fetched by each subtask will change as shards are closed and created by Kinesis.

To leverage Flink's checkpointing mechanics for exactly-once streaming processing guarantees, the Flink Kinesis consumer is implemented with the AWS Java SDK, instead of the officially recommended AWS Kinesis Client Library, for low-level control on the management of stream state. The Flink Kinesis Connector also supports setting the initial starting points of Kinesis streams, namely TRIM_HORIZON and LATEST.

Kinesis and the Flink consumer support dynamic re-sharding and shard IDs, while sequential, cannot be assumed to be consecutive. There is no perfect generic default assignment function. Default shard to subtask assignment, which is based on hash code, may result in skew, with some subtasks having many shards assigned and others none.

It is recommended to monitor the shard distribution and adjust assignment appropriately. A custom assigner implementation can be set via setShardAssigner(KinesisShardAssigner) to optimize the hash function or use static overrides to limit skew.

In order for the consumer to emit watermarks, a timestamp assigner needs to be set via setPeriodicWatermarkAssigner(AssignerWithPeriodicWatermarks) and the auto watermark emit interval configured via ExecutionConfig.setAutoWatermarkInterval(long).

Watermarks can only advance when all shards of a subtask continuously deliver records. To avoid an inactive or closed shard to block the watermark progress, the idle timeout should be configured via configuration property ConsumerConfigConstants.SHARD_IDLE_INTERVAL_MILLIS. By default, shards won't be considered idle and watermark calculation will wait for newer records to arrive from all shards.

Note that re-sharding of the Kinesis stream while an application (that relies on the Kinesis records for watermarking) is running can lead to incorrect late events. This depends on how shards are assigned to subtasks and applies regardless of whether watermarks are generated in the source or a downstream operator.

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.flink.streaming.api.functions.source.SourceFunction

org.apache.flink.streaming.api.functions.source.SourceFunction.SourceContext<T>

Constructor Summary

Constructors

Constructor and Description
FlinkKinesisConsumer(List<String> streams, KinesisDeserializationSchema<T> deserializer, Properties configProps) Creates a new Flink Kinesis Consumer.
FlinkKinesisConsumer(String stream, org.apache.flink.api.common.serialization.DeserializationSchema<T> deserializer, Properties configProps) Creates a new Flink Kinesis Consumer.
FlinkKinesisConsumer(String stream, KinesisDeserializationSchema<T> deserializer, Properties configProps) Creates a new Flink Kinesis Consumer.

Method Summary

Modifier and Type	Method and Description
void	cancel()
void	close()
protected KinesisDataFetcher<T>	createFetcher(List<String> streams, org.apache.flink.streaming.api.functions.source.SourceFunction.SourceContext<T> sourceContext, org.apache.flink.api.common.functions.RuntimeContext runtimeContext, Properties configProps, KinesisDeserializationSchema<T> deserializationSchema) This method is exposed for tests that need to mock the KinesisDataFetcher in the consumer.
org.apache.flink.streaming.api.functions.AssignerWithPeriodicWatermarks<T>	getPeriodicWatermarkAssigner()
org.apache.flink.api.common.typeinfo.TypeInformation<T>	getProducedType()
KinesisShardAssigner	getShardAssigner()
WatermarkTracker	getWatermarkTracker()
void	initializeState(org.apache.flink.runtime.state.FunctionInitializationContext context)
void	run(org.apache.flink.streaming.api.functions.source.SourceFunction.SourceContext<T> sourceContext)
void	setPeriodicWatermarkAssigner(org.apache.flink.streaming.api.functions.AssignerWithPeriodicWatermarks<T> periodicWatermarkAssigner) Set the assigner that will extract the timestamp from T and calculate the watermark.
void	setShardAssigner(KinesisShardAssigner shardAssigner) Provide a custom assigner to influence how shards are distributed over subtasks.
void	setWatermarkTracker(WatermarkTracker watermarkTracker) Set the global watermark tracker.
void	snapshotState(org.apache.flink.runtime.state.FunctionSnapshotContext context)

Methods inherited from class org.apache.flink.api.common.functions.AbstractRichFunction

getIterationRuntimeContext, getRuntimeContext, open, setRuntimeContext

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

FlinkKinesisConsumer

public FlinkKinesisConsumer(String stream,
                            org.apache.flink.api.common.serialization.DeserializationSchema<T> deserializer,
                            Properties configProps)

Creates a new Flink Kinesis Consumer.

The AWS credentials to be used, AWS region of the Kinesis streams, initial position to start streaming from are configured with a Properties instance.

Parameters:

stream - The single AWS Kinesis stream to read from.

deserializer - The deserializer used to convert raw bytes of Kinesis records to Java objects (without key).

configProps - The properties used to configure AWS credentials, AWS region, and initial starting position.

FlinkKinesisConsumer

public FlinkKinesisConsumer(String stream,
                            KinesisDeserializationSchema<T> deserializer,
                            Properties configProps)

Creates a new Flink Kinesis Consumer.

The AWS credentials to be used, AWS region of the Kinesis streams, initial position to start streaming from are configured with a Properties instance.

Parameters:

stream - The single AWS Kinesis stream to read from.

deserializer - The keyed deserializer used to convert raw bytes of Kinesis records to Java objects.

configProps - The properties used to configure AWS credentials, AWS region, and initial starting position.

FlinkKinesisConsumer

public FlinkKinesisConsumer(List<String> streams,
                            KinesisDeserializationSchema<T> deserializer,
                            Properties configProps)

Creates a new Flink Kinesis Consumer.

The AWS credentials to be used, AWS region of the Kinesis streams, initial position to start streaming from are configured with a Properties instance.

Parameters:

streams - The AWS Kinesis streams to read from.

deserializer - The keyed deserializer used to convert raw bytes of Kinesis records to Java objects.

configProps - The properties used to configure AWS credentials, AWS region, and initial starting position.

Method Detail

getShardAssigner

public KinesisShardAssigner getShardAssigner()

setShardAssigner

public void setShardAssigner(KinesisShardAssigner shardAssigner)

Provide a custom assigner to influence how shards are distributed over subtasks.

Parameters:

shardAssigner - shard assigner

getPeriodicWatermarkAssigner

public org.apache.flink.streaming.api.functions.AssignerWithPeriodicWatermarks<T> getPeriodicWatermarkAssigner()

setPeriodicWatermarkAssigner

public void setPeriodicWatermarkAssigner(org.apache.flink.streaming.api.functions.AssignerWithPeriodicWatermarks<T> periodicWatermarkAssigner)

Set the assigner that will extract the timestamp from T and calculate the watermark.

Parameters:

periodicWatermarkAssigner - periodic watermark assigner

getWatermarkTracker

public WatermarkTracker getWatermarkTracker()

setWatermarkTracker

public void setWatermarkTracker(WatermarkTracker watermarkTracker)

Set the global watermark tracker. When set, it will be used by the fetcher to align the shard consumers by event time.

Parameters:

watermarkTracker -

run

public void run(org.apache.flink.streaming.api.functions.source.SourceFunction.SourceContext<T> sourceContext)
         throws Exception

Specified by:

run in interface org.apache.flink.streaming.api.functions.source.SourceFunction<T>

Throws:

Exception

cancel

public void cancel()

Specified by:

cancel in interface org.apache.flink.streaming.api.functions.source.SourceFunction<T>

close

public void close()
           throws Exception

Specified by:

close in interface org.apache.flink.api.common.functions.RichFunction

Overrides:

close in class org.apache.flink.api.common.functions.AbstractRichFunction

Throws:

Exception

getProducedType

public org.apache.flink.api.common.typeinfo.TypeInformation<T> getProducedType()

Specified by:

getProducedType in interface org.apache.flink.api.java.typeutils.ResultTypeQueryable<T>

initializeState

public void initializeState(org.apache.flink.runtime.state.FunctionInitializationContext context)
                     throws Exception

Specified by:

initializeState in interface org.apache.flink.streaming.api.checkpoint.CheckpointedFunction

Throws:

Exception

snapshotState

public void snapshotState(org.apache.flink.runtime.state.FunctionSnapshotContext context)
                   throws Exception

Specified by:

snapshotState in interface org.apache.flink.streaming.api.checkpoint.CheckpointedFunction

Throws:

Exception

createFetcher

protected KinesisDataFetcher<T> createFetcher(List<String> streams,
                                              org.apache.flink.streaming.api.functions.source.SourceFunction.SourceContext<T> sourceContext,
                                              org.apache.flink.api.common.functions.RuntimeContext runtimeContext,
                                              Properties configProps,
                                              KinesisDeserializationSchema<T> deserializationSchema)

This method is exposed for tests that need to mock the KinesisDataFetcher in the consumer.