[FLINK-38606][Connectors/MongoDB] Support for Flink 2.1.0 API

[neoneo008]

Below changes are required for support flink version 2.1.0

Mongo Source:
API change for SingleThreadFetcherManager constructor to remove FutureCompletingBlockingQueue. This is now created internally by SingeThreadFetcherManager using Configuration for queue capacity.

Mongo Sink:
Migrate from InitContext to WriterInitContext (API change in flink 2.x)

Both of the above changes are driven from API changes well documented on flink release notes.

[boring-cyborg]

Thanks for opening this pull request! Please check out our contributing guidelines. (https://flink.apache.org/contributing/how-to-contribute.html)

[Jiabao-Sun]

Thank you for your contribution, @neoneo008 .
This looks good overall, but I think we should preserve JDK 11 support for now.

[neoneo008]

The test case failures are due to a Flink bug https://issues.apache.org/jira/browse/FLINK-37546 . They kyro scala initialization is required to support Scala API.

Will need to wait until the original issue is resolved.

[davidradl]

I would look to the way that the Kafka connector does these github actions; i.e. the push_pr should check against releases not snapshot levels of Flink
Namely:

  compile_and_test:
    strategy:
      matrix:
        flink: &flink_versions [ 2.0.1, 2.1.1 ]
        jdk: &jdk_versions [ '11, 17, 21' ]

Then amend the weekly github action in the style of the Kafka connector including adding inflink: 2.2-SNAPSHOT, to check it works with the latest Flink

We have also made a similar change to the JDBC connector recently. fyi @ferenc-csaky