** Hazelcast Jet 3.1 **

* Apache Beam Runner for Hazelcast Jet is released. For more details please see https://beam.apache.org/documentation/runners/jet/
* New hazelcast-jet-contrib repository containing new connectors:
   * InfluxDb: https://github.com/hazelcast/hazelcast-jet-contrib/tree/master/influxdb
   * Elasticsearch: https://github.com/hazelcast/hazelcast-jet-contrib/tree/master/elasticsearch
* New AggregateOperations: `filtering` and `flatMapping`
* SourceBuilder API has been extended to add support for fault tolerance
* Support for composite transforms in Pipeline API through the `apply` operator
* Jet now supports YAML configuration as well as XML
* Update Hazelcast IMDG version to 3.12.1

** Hazelcast Jet 3.0 **

* 3.0 is the first major release of Hazelcast Jet past 0.x, and is versioned at 3.x to align
it better with IMDG 3.x versions that it is based on
* Starting with 3.0, main Job API and the pipeline API will be backwards compatible in further minor
versions. Core API will only be compatible between patch versions.
  - Several APIs have been streamlined in preparation for this change. Please see the breaking changes
  section of the release notes for a list of specific changes.
* Pipeline API has the following new additions:
  - Support for async map/filter/flatMapUsingContext methods
  - Support for emitting early results for a window before a window has completed
* Being able to atomically create a job only if it's not already running using
  JetInstance.newJobIfAbsent
* [Jet Enterprise] Support for job upgrades and A/B testing via taking a named snapshot and then starting the job
  from the saved snapshot.
* [Jet Enterprise] Being able to restart the whole cluster and restoring jobs by using
  IMDG Hot Restart to persist the cluster to disk
* New Jet Command Line Interface for job management of a cluster
* Support for improved Watermark diagnostics in Jet Management Center
* Support for showing map stats in Jet Management Center
* Full support for JDK 11
* Update Hazelcast version to 3.12

** Hazelcast Jet 0.7.2 **

* Fix for missing config schema file in previous release

** Hazelcast Jet 0.7.1 **

* Full support for running Hazelcast Jet on Kubernetes
* Hazelcast Jet Helm Charts are released
* TopN and BottomN aggregate operations are added

** Hazelcast Jet 0.7 **

* Several new additions to pipeline API including:
  - A new SourceBuilder API which can be used to build both streaming and batch sources
  - Merge and distinct operations
  - Simplified ways to do aggregations on more than one input (co-aggregation)
  - Support for rolling aggregations
  - Keyed map/flatMap/filterUsingContext operations
  - Convenience functions for enriching a stream directly from an IMap or ReplicatedMap
* Several additions to job management:
  - Jobs can now be automatically both scaled up and down without disruption when new nodes are added or removed.
  - Jobs can now be suspended and resumed.
* Several new default connectors are added:
  - JMS
  - JDBC
  - Avro Files
* Distributed java.util.stream implementation has been removed with this release.
  Please refer to the Pipeline API instead.
* It's now possible to get Jet metrics through a JMX interface.
* First version of Jet Management Center is released
* Update Hazelcast version to 3.10.5
* Security suite available in Jet enterprise version

** Hazelcast Jet 0.6.1 **

* Optimize in memory layout of snapshot data, reducing memory usage by up to 75%
* Add support for configuring a custom classloader per job
* Update Hazelcast version to final released version (3.10)

** Hazelcast Jet 0.6 **

* Pipeline API has been extended with support for windowing operations
including co grouped windows and support for event time.
* Pipeline API now allows stateful transforms using a context object.
* Parallelism and name of each stage on the pipeline API can now be set individually.
* New custom sink builder for building sink connectors without writing custom processors.
* Merging and updating map sinks, which can update a value in place rather than overwriting it.
* It is now possible to set names for jobs and retrieve them by name.
* Running jobs can now be manually restarted and scaled up.
* Spring integration module which brings support for XML and annotation
based Spring configuration
* Jet builds on Java 9 and has an explicit automatic module name.
* New Jet demos repository added to showcase several new use cases for Jet.
* Hazelcast version is updated to 3.10.

** Hazelcast Jet 0.5 **

* Introduction of Pipeline API for building Jet applications with support for
  group by, hash join and co-group operations. The DAG API is still supported,
  and referred to as the core API.
* Generalization of AggregateOperation abstraction to support multiple inbound
  streams.
* Improved job lifecycle management - jobs can now be restarted automatically
  when a node leaves the cluster.
* Exactly-once and at-least-once processing guarantees when restarting streaming
  jobs with use of distributed in memory snapshotting.
* New streaming source from IMap and ICache using Hazelcast event journal
* Support for predicates and projections when reading from Hazelcast IMap
* Ability to create a distributed java.util.stream from any source.
* Update Hazelcast version to 3.9.

** Hazelcast Jet 0.4 **

* Improved streaming support including windowing support with event-time semantics.
* Out of the box support for tumbling, sliding and session window aggregations.
* Streaming file reader to process change streams of files and general improvements
  to file readers and writers.
* New AggregateOperation abstraction with several built-in ones including count,
  average, sum, min, max and linear regression.
* Hazelcast ICache support added. ICache can be used as a source or sink with both
  core API and java.util.stream.
* Socket connector is overhauled and can now also be used as a sink.
* Many new code samples including several streaming examples and enrichment and
  co-group for batch operations.
* Hazelcast version updated to 3.8.2. Hazelcast IMDG is now shaded inside the
  Jet JAR.
* Several built-in diagnostic processors and unit test support for writing custom
  Processors.
* Hazelcast Jet tile for Pivotal Cloud Foundry is released,

** Hazelcast Jet 0.3.1 **

* The Jet cluster can be scaled up without affecting running jobs.
  Please refer to the [elasticity](https://github.com/hazelcast/hazelcast-jet-reference-manual/blob/master/src/Introduction.md#elasticity)
  section in the user guide for details.
* Added support for reading and writing non-text data from or to HDFS.
  Please refer to the  [hazelcast-jet-hadoop documentation](https://github.com/hazelcast/hazelcast-jet-reference-manual/blob/master/src/Modules.md#hazelcast-jet-hadoop) for details.
* Kafka Connector extended. Please refer to the [hazelcast-jet-kafka documentation](https://github.com/hazelcast/hazelcast-jet-reference-manual/blob/master/src/Modules.md#hazelcast-jet-kafka) for details.
* Hazelcast version updated to 3.8
