[Task]: Beam Flink pipeline engine page: Flink version, contradicting job id sections and Javadoc leftovers #9048

What needs to happen?

pipeline/pipeline-run-configurations/beam-flink-pipeline-engine.adoc has a few outdated or contradicting parts:

• "This runner allows you to run Hop pipelines on Apache Flink version 1.13." The Beam plugin uses beam-runners-flink-1.19, and the compatibility table in pipeline/beam/getting-started-with-beam.adoc says Flink 1.19.x.
• Two sections disagree about the Flink job id, and each is only right for some Flink masters. "Job id" (added with Issue #3131 : Expose the submitted Flink job id on the pipeline engine #8679) says Hop chooses the id before submitting and exposes it with getFlinkJobId() and the flink.job.id execution detail. "Flink job ID" at the end of the page (added with Issue #8708 : Add Beam MQTT, Debezium, Elasticsearch, Splunk and Snowflake IO #8716) says "Hop does not show a Flink job ID after a Beam run". In BeamFlinkPipelineEngine, assignFlinkJobId() only fixes a job id when the Flink master is [local] or host:port. With [auto], which is the default, with an empty master, or with [collection], no id is assigned: with [auto] the environment comes from flink run or the Kubernetes operator. The two sections could be merged into one: Hop assigns and reports the id for a [local] or host:port master; with [auto] or [collection] there is no id, and the job can be found by its Beam job name.
• The options table has Javadoc leftovers: "A value of {@code -1}" (Execution retry delay) and "Reference {@link org.apache.flink.api.common.ExecutionMode}" (Batch execution mode).

Issue Priority

Priority: 3

Issue Component

Component: Documentation