[FLINK-39132] [e2e] Stabilize Kubernetes materialized table test

[qiuyanjun888]

What is the purpose of the change

This pull request stabilizes the Kubernetes Materialized Table end-to-end test for FLINK-39132.

The test suspends a continuous materialized table by stopping the Kubernetes application job with a savepoint. The Jira evidence shows that this stop-with-savepoint operation can exceed the default 60s SQL Gateway client timeout. A later linked Azure failure also shows the same test failing during post-test log validation because Hadoop S3A emits a harmless statistics field named stream_write_exceptions=0, which the generic e2e exception grep treats as an exception.

Brief change log

• Configure the Kubernetes Materialized Table SQL Gateway session with client.timeout = 5min so the stop-with-savepoint operation has enough time in CI.
• Use a custom log-check action for this test that still checks errors, real exceptions, and non-empty .out files, but allows the zero-valued S3A stream_write_exceptions=0 statistics field.

Verifying this change

This change added targeted validation for the e2e shell-script behavior and can be verified as follows:

• bash -n flink-end-to-end-tests/test-scripts/test_kubernetes_materialized_table.sh
• bash -n flink-end-to-end-tests/run-nightly-tests.sh
• A focused shell harness verified that test_kubernetes_materialized_table.sh check_exceptions passes for the S3A stream_write_exceptions=0 statistics line.
• The same harness verified that test_kubernetes_materialized_table.sh check_exceptions still fails for a real java.lang.RuntimeException log line.
• A focused static check verified that the test session options include client.timeout = 5min.
• git diff --check

I did not run the full Kubernetes e2e test locally because it requires the full Flink distribution plus minikube/kubectl/sudo setup used by CI.

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

• Dependencies (does it add or upgrade a dependency): no
• The public API, i.e., is any changed class annotated with @Public(Evolving): no
• The serializers: no
• The runtime per-record code paths (performance sensitive): no
• Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: yes, Kubernetes e2e test configuration only
• The S3 file system connector: no, only the e2e log checker allow-list for a zero-valued S3A metric field

Documentation

• Does this pull request introduce a new feature? no

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Hermes Agent

[flinkbot]

CI report:

• 2733591 Azure: FAILURE

Bot commands

The @flinkbot bot supports the following commands:

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

[qiuyanjun888]

This PR is focused on end-to-end Kubernetes materialized table test.

@RocMarshal @MartijnVisser could you please take a look when you have time and advise whether this approach is acceptable for this area and can move forward?