TRON-2277: Pass along non_retryable_exit_codes to KubernetesCluster objects

[jfongatyelp]

Looks like we missed one place where we need to pass along the k8s_options from MASTER configs, when we instantiate our KubernetesClusters from KubernetesClusterRepository.

Applying this change to a local instance and infrastage now do what we expect and do not retry failed actions if their exit codes are in non_retryable_exit_codes, plus logs what we expect:

2024-09-25 17:00:22,786 tron.core.actionrun INFO ActionRun: MASTER.testjob0.14.zeroth skipping auto-retries, received non-retryable exit code (72).
2024-09-25 17:00:22,786 tron.core.actionrun INFO ActionRun: MASTER.testjob0.14.zeroth completed with fail, transitioned to running, exit status: 72
2024-09-25 17:00:22,786 tron.utils.state DEBUG transitioning from running to failed
2024-09-25 17:00:22,786 tron.utils.observer DEBUG Notifying 1 listeners for new event 'failed'
2024-09-25 17:00:22,786 tron.core.jobrun INFO JobRun:MASTER.testjob0.14 got an event: failed

(i.e. no retries after the non-retryable exit code line + goes to jobrun failed)

This also adds 2 fairly dumb tests (the actionrun test alone wouldn't have caught the fix since we're mocking get_cluster's result itself, hence why I also added a really dumb KubernetesClusterRepository->KubernetesCluster one to make sure we don't skip passing along new configs here in the future.

Let me know if you have a better idea for any tests!

[nemacysts]

good catch! maybe one day we'll reduce some of the layers here :(

and ++ for the new options test!