In general, you package your DataStream, Table API or DataSet API program for Ververica Platform in the same as for any other Apache Flink® runtime.
The Apache Flink® documentation provides example project setups for Maven & Gradle to get you started.
Ververica Platform comes with our own distribution of Apache Flink®.
Our distribution is always compatible with vanilla Apache Flink®, but might include additional and/or more timely bug and security vulnerability fixes.
Our Apache Flink® Docker images are built from our distribution of Apache Flink®, but the Docker images do not contain all Apache Flink® dependencies that you might need to run your program.
Most importantly, connector dependencies, e.g. for Apache Kafka®, are not provided by the Ververica runtime, but need to be bundled by the user.
In order for you to use our distribution of Apache Flink® when packaging your program, we host a Maven repository including these dependencies.
Please follow the instructions to integrate it with your build.
Second, our distribution of Apache Flink® is published under a different Maven group id and version than vanilla Apache Flink®.
The group id is “com.ververica.flink” instead of “org.apache.flink”.
The versions follows the versioning of Apache Flink® with an additional counter for off-cycle releases: “<major>.`<minor>`.`<patch>`-<stream|spring>``<ververica-version-counter>”.
For example, our second releases for Apache Flink® 1.11.2 would have the versions “1.11.2-stream2”” and “1.11.2-spring2”.
The stream|spring tag identifies the edition of Ververica Platform that you are using.