[FLINK-38403] Stabilize UnalignedCheckpoint ITs

[AHeise]

What is the purpose of the change

To take a checkpoint, UC ITs relied on a limited number of retries after failing artificially to produce a checkpoint with in-flight data. Naturally, this approach is flaky when the actual test environment is flaky and retry attempts are used up quickly for unplanned restarts.

The solution is to stop fiddling around with the number of retries and instead use NonRecoverableError to force a failure of the job independent of the retries.

This commit also switches to a better way to retrieve the path of that checkpoint (through JobManager API).

Brief change log

• Improves stability of UnalignedCheckpoint ITs

Verifying this change

Test only change

Does this pull request potentially affect one of the following parts:

• Dependencies (does it add or upgrade a dependency): (yes / no)
• The public API, i.e., is any changed class annotated with @Public(Evolving): (yes / no)
• The serializers: (yes / no / don't know)
• The runtime per-record code paths (performance sensitive): (yes / no / don't know)
• Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: (yes / no / don't know)
• The S3 file system connector: (yes / no / don't know)

Documentation

• Does this pull request introduce a new feature? (yes / no)
• If yes, how is the feature documented? (not applicable / docs / JavaDocs / not documented)

[flinkbot]

CI report:

• 7c01916 Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot run azure re-run the last Azure build

[Zakelly]

Thanks for the fix! LGTM.