classicvalues / amazon-kinesis-client

Client library for Amazon Kinesis

Amazon Kinesis Client Library for Java

The Amazon Kinesis Client Library for Java (Amazon KCL) enables Java developers to easily consume and process data from Amazon Kinesis.

• Kinesis Product Page
• Forum
• Issues

Features

• Provides an easy-to-use programming model for processing data using Amazon Kinesis
• Helps with scale-out and fault-tolerant processing

Getting Started

1. Sign up for AWS — Before you begin, you need an AWS account. For more information about creating an AWS account and retrieving your AWS credentials, see AWS Account and Credentials in the AWS SDK for Java Developer Guide.
2. Sign up for Amazon Kinesis — Go to the Amazon Kinesis console to sign up for the service and create an Amazon Kinesis stream. For more information, see Create an Amazon Kinesis Stream in the Amazon Kinesis Developer Guide.
3. Minimum requirements — To use the Amazon Kinesis Client Library, you'll need Java 1.8+. For more information about Amazon Kinesis Client Library requirements, see Before You Begin in the Amazon Kinesis Developer Guide.
4. Using the Amazon Kinesis Client Library — The best way to get familiar with the Amazon Kinesis Client Library is to read Developing Record Consumer Applications in the Amazon Kinesis Developer Guide.

Building from Source

After you've downloaded the code from GitHub, you can build it using Maven. To disable GPG signing in the build, use this command: mvn clean install -Dgpg.skip=true

Integration with the Kinesis Producer Library

For producer-side developers using the Kinesis Producer Library (KPL), the KCL integrates without additional effort. When the KCL retrieves an aggregated Amazon Kinesis record consisting of multiple KPL user records, it will automatically invoke the KPL to extract the individual user records before returning them to the user.

Amazon KCL support for other languages

To make it easier for developers to write record processors in other languages, we have implemented a Java based daemon, called MultiLangDaemon that does all the heavy lifting. Our approach has the daemon spawn a sub-process, which in turn runs the record processor, which can be written in any language. The MultiLangDaemon process and the record processor sub-process communicate with each other over STDIN and STDOUT using a defined protocol. There will be a one to one correspondence amongst record processors, child processes, and shards. For Python developers specifically, we have abstracted these implementation details away and expose an interface that enables you to focus on writing record processing logic in Python. This approach enables KCL to be language agnostic, while providing identical features and similar parallel processing model across all languages.

Release Notes

Release 1.8.8

• Fixed issues with leases losses due to ExpiredIteratorException in PrefetchGetRecordsCache and AsynchronousFetchingStrategy.
  PrefetchGetRecordsCache will request for a new iterator and start fetching data again.

  • PR#263

• Added warning message for long running tasks.
  Logging long running tasks can be enabled by setting the following configuration property:

  Name	Default	Description
  logWarningForTaskAfterMillis	Not set	Milliseconds after which the logger will log a warning message for the long running task

  • PR#259

• Handling spurious lease renewal failures gracefully.
  Added better handling of DynamoDB failures when updating leases. These failures would occur when a request to DynamoDB appeared to fail, but was actually successful.

  • PR#247

• ShutdownTask gets retried if the previous attempt on the ShutdownTask fails.

  • PR#267

• Fix for using maxRecords from KinesisClientLibConfiguration in GetRecordsCache for fetching records.

  • PR#264

Release 1.8.7

• Don't add a delay for synchronous requests to Kinesis
  Removes a delay that had been added for synchronous GetRecords calls to Kinesis.
  • PR #256

Release 1.8.6

• Add prefetching of records from Kinesis
  Prefetching will retrieve and queue additional records from Kinesis while the application is processing existing records.
  Prefetching can be enabled by setting dataFetchingStrategy to PREFETCH_CACHED. Once enabled an additional fetching thread will be started to retrieve records from Kinesis. Retrieved records will be held in a queue until the application is ready to process them.
  Pre-fetching supports the following configuration values:

  Name	Default	Description
  dataFetchingStrategy	DEFAULT	Which data fetching strategy to use
  maxPendingProcessRecordsInput	3	The maximum number of process records input that can be queued
  maxCacheByteSize	8 MiB	The maximum number of bytes that can be queued
  maxRecordsCount	30,000	The maximum number of records that can be queued
  idleMillisBetweenCalls	1,500 ms	The amount of time to wait between calls to Kinesis

  • PR #246

Release 1.8.5 (September 26, 2017)

• Only advance the shard iterator for the accepted response.
  This fixes a race condition in the KinesisDataFetcher when it's being used to make asynchronous requests. The shard iterator is now only advanced when the retriever calls DataFetcherResult#accept().
  • PR #230
  • Issue #231

Release 1.8.4 (September 22, 2017)

• Create a new completion service for each request.
  This ensures that canceled tasks are discarded. This will prevent a cancellation exception causing issues processing records.
  • PR #227
  • Issue #226

Release 1.8.3 (September 22, 2017)

• Call shutdown on the retriever when the record processor is being shutdown
  This fixes a bug that could leak threads if using the AsynchronousGetRecordsRetrievalStrategy is being used.
  The asynchronous retriever is only used when KinesisClientLibConfiguration#retryGetRecordsInSeconds, and KinesisClientLibConfiguration#maxGetRecordsThreadPool are set.
  • PR #222

Release 1.8.2 (September 20, 2017)

• Add support for two phase checkpoints
  Applications can now set a pending checkpoint, before completing the checkpoint operation. Once the application has completed its checkpoint steps, the final checkpoint will clear the pending checkpoint.
  Should the checkpoint fail the attempted sequence number is provided in the InitializationInput#getPendingCheckpointSequenceNumber otherwise the value will be null.
  • PR #188
• Support timeouts, and retry for GetRecords calls.
  Applications can now set timeouts for GetRecord calls to Kinesis. As part of setting the timeout, the application must also provide a thread pool size for concurrent requests.
  • PR #214
• Notification when the lease table is throttled
  When writes, or reads, to the lease table are throttled a warning will be emitted. If you're seeing this warning you should increase the IOPs for your lease table to prevent processing delays.
  • PR #212
• Support configuring the graceful shutdown timeout for MultiLang Clients
  This adds support for setting the timeout that the Java process will wait for the MutliLang client to complete graceful shutdown. The timeout can be configured by adding shutdownGraceMillis to the properties file set to the number of milliseconds to wait.
  • PR #204

Release 1.8.1 (August 2, 2017)

• Support timeouts for calls to the MultiLang Daemon This adds support for setting a timeout when dispatching records to the client record processor. If the record processor doesn't respond within the timeout the parent Java process will be terminated. This is a temporary fix to handle cases where the KCL becomes blocked while waiting for a client record processor. The timeout for the this can be set by adding timeoutInSeconds = <timeout value>. The default for this is no timeout.
  Setting this can cause the KCL to exit suddenly, before using this ensure that you have an automated restart for your application
  • PR #195
  • Issue #185

Release 1.8.0 (July 25, 2017)

• Execute graceful shutdown on its own thread
  • PR #191
  • Issue #167
• Added support for controlling the size of the lease renewer thread pool
  • PR #177
  • Issue #171
• Require Java 8 and later
  Java 8 is now required for versions 1.8.0 of the amazon-kinesis-client and later.
  • PR #176

Release 1.7.6 (June 21, 2017)

• Added support for graceful shutdown in MultiLang Clients
  • PR #174
  • PR #182
• Updated documentation for v2.IRecordProcessor#shutdown, and KinesisClientLibConfiguration#idleTimeBetweenReadsMillis
  • PR #170
• Updated to version 1.11.151 of the AWS Java SDK
  • PR #183

Release 1.7.5 (April 7, 2017)

• Correctly handle throttling for DescribeStream, and save accumulated progress from individual calls.
  • PR #152
• Upgrade to version 1.11.115 of the AWS Java SDK
  • PR #155

Release 1.7.4 (February 27, 2017)

• Fixed an issue building JavaDoc for Java 8.
  • Issue #18
  • PR #141
• Reduce Throttling Messages to WARN, unless throttling occurs 6 times consecutively.
  • Issue #4
  • PR #140
• Fixed two bugs occurring in requestShutdown.
  • Fixed a bug that prevented the worker from shutting down, via requestShutdown, when no leases were held.
    • Issue #128
  • Fixed a bug that could trigger a NullPointerException if leases changed during requestShutdown.
    • Issue #129
  • PR #139
• Upgraded the AWS SDK Version to 1.11.91
  • PR #138
• Use an executor returned from ExecutorService.newFixedThreadPool instead of constructing it by hand.
  • PR #135
• Correctly initialize DynamoDB client, when endpoint is explicitly set.
  • PR #142

Release 1.7.3 (January 9, 2017)

• Upgrade to the newest AWS Java SDK.
  • Amazon Kinesis Client Issue #27
  • PR #126
  • PR #125
• Added a direct dependency on commons-logging.
  • Issue #123
  • PR #124
• Make ShardInfo public to allow for custom ShardPrioritization strategies.
  • Issue #120
  • PR #127

Release 1.7.2 (November 7, 2016)

• MultiLangDaemon Feature Updates The MultiLangDaemon has been upgraded to use the v2 interfaces, which allows access to enhanced checkpointing, and more information during record processor initialization. The MultiLangDaemon clients must be updated before they can take advantage of these new features.

Release 1.7.1 (November 3, 2016)

• General
  • Allow disabling shard synchronization at startup.
    • Applications can disable shard synchronization at startup. Disabling shard synchronization can application startup times for very large streams.
    • PR #102
  • Applications can now request a graceful shutdown, and record processors that implement the IShutdownNotificationAware will be given a chance to checkpoint before being shutdown.
    • This adds a new interface, and a new method on Worker.
    • PR #109
    • Solves Issue #79
• MultiLangDaemon
  • Applications can now use credential provides that accept string parameters.
    • PR #99
  • Applications can now use different credentials for each service.
    • PR #111

Release 1.7.0 (August 22, 2016)

• Add support for time based iterators (See GetShardIterator Documentation)
  • PR #94 The KinesisClientLibConfiguration now supports providing an initial time stamp position.
  • This position is only used if there is no current checkpoint for the shard.
  • This setting cannot be used with DynamoDB Streams Resolves Issue #88
• Allow Prioritization of Parent Shards for Task Assignment
  • PR #95 The KinesisClientLibconfiguration now supports providing a ShardPrioritization strategy. This strategy controls how the Worker determines which ShardConsumer to call next. This can improve processing for streams that split often, such as DynamoDB Streams.
• Remove direct dependency on aws-java-sdk-core, to allow independent versioning.
  • PR #92 You may need to add a direct dependency on aws-java-sdk-core if other dependencies include an older version.

Release 1.6.5 (July 25, 2016)

• Change LeaseManager to call DescribeTable before attempting to create the lease table.
  • Issue #36
  • PR #41
  • PR #67
• Allow DynamoDB lease table name to be specified
  • PR #61
• Add approximateArrivalTimestamp for JsonFriendlyRecord
  • PR #86
• Shutdown lease renewal thread pool on exit.
  • PR #84
• Wait for CloudWatch publishing thread to finish before exiting.
  • PR #82
• Added unit, and integration tests for the library.

Release 1.6.4 (July 6, 2016)

• Upgrade to AWS SDK for Java 1.11.14
  • Issue #74
  • Issue #73
• Maven Artifact Signing Change
  • Artifacts are now signed by the identity Amazon Kinesis Tools <[email protected]>

Release 1.6.3 (May 12, 2016)

• Fix format exception caused by DEBUG log in LeaseTaker Issue # 68

Release 1.6.2 (March 23, 2016)

• Support for specifying max leases per worker and max leases to steal at a time.
• Support for specifying initial DynamoDB table read and write capacity.
• Support for parallel lease renewal.
• Support for graceful worker shutdown.
• Change DefaultCWMetricsPublisher log level to debug. PR # 49
• Avoid NPE in MLD record processor shutdown if record processor was not initialized. Issue # 29

Release 1.6.1 (September 23, 2015)

• Expose approximateArrivalTimestamp for Records in processRecords API call.

Release 1.6.0 (July 31, 2015)

• Restores compatibility with dynamodb-streams-kinesis-adapter (which was broken in 1.4.0).

Release 1.5.1 (July 20, 2015)

• KCL maven artifact 1.5.0 does not work with JDK 7. This release addresses this issue.

Release 1.5.0 (July 9, 2015)

• Metrics Enhancements
  • Support metrics level and dimension configurations to control CloudWatch metrics emitted by the KCL.
  • Add new metrics that track time spent in record processor methods.
  • Disable WorkerIdentifier dimension by default.
• Exception Reporting — Do not silently ignore exceptions in ShardConsumer.
• AWS SDK Component Dependencies — Depend only on AWS SDK components that are used.

Release 1.4.0 (June 2, 2015)

• Integration with the Kinesis Producer Library (KPL)
  • Automatically de-aggregate records put into the Kinesis stream using the KPL.
  • Support checkpointing at the individual user record level when multiple user records are aggregated into one Kinesis record using the KPL.

See Consumer De-aggregation with the KCL for details.

Release 1.3.0 (May 22, 2015)

• A new metric called "MillisBehindLatest", which tracks how far consumers are from real time, is now uploaded to CloudWatch.

Release 1.2.1 (January 26, 2015)

• MultiLangDaemon — Changes to the MultiLangDaemon to make it easier to provide a custom worker.

Release 1.2 (October 21, 2014)

• Multi-Language Support — Amazon KCL now supports implementing record processors in any language by communicating with the daemon over STDIN and STDOUT. Python developers can directly use the Amazon Kinesis Client Library for Python to write their data processing applications.

Release 1.1 (June 30, 2014)

• Checkpointing at a specific sequence number — The IRecordProcessorCheckpointer interface now supports checkpointing at a sequence number specified by the record processor.
• Set region — KinesisClientLibConfiguration now supports setting the region name to indicate the location of the Amazon Kinesis service. The Amazon DynamoDB table and Amazon CloudWatch metrics associated with your application will also use this region setting.