Skip to end of metadata
Go to start of metadata

This document is a community maintained roadmap of the upcoming releases and features (at a high level).
The document contains only estimations, no guaranteed feature / release dates. 

See also the time-based release model.


Last update: May 2016

Next scheduled major release: 1.4.0
Feature freeze (branch forking):  4. September 2017
Code freeze (first voting RC):  18 September 2017
Release date29 September 2017


ReleaseDateTime span (days)



26 Aug 2014 
0.704 Nov 201470
0.821 Jan 201578
0.924 Jun 2015154
0.1016 Nov 2015145
1.008 Mar 2016113
1.108 Aug 2016153


2 Feb 2017178
1.326 May 2017119
1.429 September 2017 
1.526 January 2018 


Past releases

Flink 1.1 (August 2016)

  • Session Windows and Allowed lateness
  • Table API and SQL on Calcite
  • Metrics system
  • Scala API for CEP
  • Graph generators and new Gelly library algorithms
  • Continuous File Sources
  • Kinesis Connector
  • Cassandra Connector

Flink 1.2

Flink 1.3

  • (Large) State Handling / Recovery:

  • Streaming APIs:

  • Deployment & Tooling:

    • Flink History Server (FLINK-1579)

    • Flink automatically manages TM network buffer (FLINK-4545)

    • Add Datadog HTTP metrics reporter (FLINK-6013)

    • Add watermark monitoring to JobManager web frontend (FLINK-3427)

  • Table API / SQL:

    • Support for Retraction in Table API / SQL (FLINK-6047)

    • Support for non-windowed aggregations in streaming Table API / SQL (FLINK-6216)
    • Support for append, upsert and retraction StreamTableSinks (FLINK-6093)
    • Integrate time indicators for Table API & SQL (FLINK-5884)

    • Support for group windows in (batch + streaming) SQL (TUMBLE, HOP, SESSION) (FLINK-6261,FLINK-6011)
    • Support for OVER windows in streaming Table API and SQL (FLINK-4557)
    • Support for QueryConfig for streaming queries to specify state retention interval (FLINK-6491)
    • Support for user-defined aggregation functions (FLINK-5564)
    • Support for standard deviation and variance aggregation functions (FLINK-4604)
    • Support for TableSources with filter push down (FLINK-3849,

    • Support for external catalogs (FLINK-5570)

    • Add KafkaAvroTableSource (FLINK-3871)
    • Add HBaseTableSource (FLINK-2168)

    • Add CUBE/ROLLUP/GROUPING SETS operator in batch SQL (FLINK-5303)

    • Support for DISTINCT GROUP BY aggregates for batch SQL (FLINK-3849)
    • Add flink-table.jar to ./opt folder (FLINK-6247)
    • Support for more scalar functions (SIN, COS, TAN, DEGREES, PI etc.) (FLINK-6112)
  • Connectors:
    • Elasticsearch 5.x support (FLINK-4988)

    • Implement rescalable non-partitioned state for Kinesis Connector (FLINK-4821)

    • Transparent shard discovery for Kinesis Consumer (FLINK-4577)
    • Allow Kinesis Consumer to start from specific timestamp / Date (FLINK-4523)

    • Allow setting custom start-offsets for the Kafka consumer (FLINK-3123)

    • Allow out-opt from offset committing for the Kafka consumer (FLINK-3398)

  • CEP library improvements:

  • Gelly Library:
    • Unified driver for running Gelly examples (FLINK-4949)
    • PageRank algorithm for directed graphs (FLINK-4896)

    • Add Circulant and Echo graph generators (FLINK-6393)

Upcoming releases

Flink 1.4

Flink 1.5




  • No labels