Add embedded test for streaming logs of cancelled task

[kfaraz]

Description

Changes:

• Add embedded test to verify streaming logs of a cancelled task in all environments

---

This PR has:

• been self-reviewed.
  • using the concurrency checklist (Remove this item if the PR doesn't have any relation to concurrency.)
• added documentation for new or modified features or behaviors.
• a release note entry in the PR description.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.

[capistrant]

+1 pending green CI!

[kfaraz]

The KubernetesWithClusterOperatorDockerTest failed if we remove the invocation of startWatchingLogs() from KubernetesWorkItem.shutdown().

https://github.com/apache/druid/actions/runs/18058444959/job/51392286069?pr=18580#step:9:8155

Trying the test once with the original code.

[kfaraz]

The KubernetesClusterWithOperatorDockerTest succeeded with the original code.

@capistrant , it seems that we can't get rid of the invocation of startWatchingLogs() in KubernetesWorkItem.shutdown().
I haven't dug very deep but it is most likely because
KubernetesWorkItem.shutdown() calls KubernetesPeonLifecycle.shutdown()
which calls kubernetesClient.deletePeonJob(taskId).

After this, join() calls startWatchingLogs() but it fails to initialize logWatch since the peon job is already deleted.

[kfaraz]

I have triggered another run to confirm this.

[capistrant]

The KubernetesClusterWithOperatorDockerTest succeeded with the original code.

@capistrant , it seems that we can't get rid of the invocation of startWatchingLogs() in KubernetesWorkItem.shutdown(). I haven't dug very deep but it is most likely because KubernetesWorkItem.shutdown() calls KubernetesPeonLifecycle.shutdown() which calls kubernetesClient.deletePeonJob(taskId).

After this, join() calls startWatchingLogs() but it fails to initialize logWatch since the peon job is already deleted.

Damn ok. #18556 did upgrade fabric8 to 7.4.0 which attempts to address the bug that is seemingly behind the LogWatch issues (fabric8io/kubernetes-client#7163) so maybe it isn't huge deal if this bug is actually fixed in fabric8

[kfaraz]

Damn ok. #18556 did upgrade fabric8 to 7.4.0 which attempts to address the bug that is seemingly behind the LogWatch issues (fabric8io/kubernetes-client#7163) so maybe it isn't huge deal if this bug is actually fixed in fabric8

Yeah, let's hope the issue doesn't happen after the upgrade.

I will merge this PR off once the CI is green since the test is still useful.

The embedded test flakiness encountered here is being addressed in #18583