[ci][RLlib] [test utils] Improve error message on some test-failures because of IndexError in test_utils.py

[Daraan]

Why are these changes needed?

In the CI I saw a fleaky test failing, but its error is not reported correctly:

[2025-07-04T00:01:57Z] Traceback (most recent call last):
[2025-07-04T00:01:57Z]   File "/root/.cache/bazel/_bazel_root/1df605deb6d24fc8068f6e25793ec703/execroot/com_github_ray_project_ray/bazel-out/k8-opt/bin/rllib/examples/connectors/multi_agent_observation_preprocessor.runfiles/com_github_ray_project_ray/rllib/examples/connectors/multi_agent_observation_preprocessor.py", line 137, in <module>
[2025-07-04T00:01:57Z]     run_rllib_example_script_experiment(base_config, args)
[2025-07-04T00:01:57Z]   File "/rayci/python/ray/rllib/utils/test_utils.py", line 1354, in run_rllib_example_script_experiment
[2025-07-04T00:01:57Z]     f"{[e.args[0].args[2] for e in results.errors]}"
[2025-07-04T00:01:57Z]   File "/rayci/python/ray/rllib/utils/test_utils.py", line 1354, in <listcomp>
[2025-07-04T00:01:57Z]     f"{[e.args[0].args[2] for e in results.errors]}"  # <---
[2025-07-04T00:01:57Z] IndexError: tuple index out of range  # <---

In the logs the actual error does not appear because of the IndexError for f"{[e.args[0].args[2] for e in results.errors]}". I am not sure why e.args[0].args[2] is used, a nested Exception perhaps? This PR adds a guard to fallback to a simpler repr(e) in case the IndexError would accur again. This allows to investigate why this IndexError occures and raises the RuntimeError like expected.

Related issue number

NA

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • This PR is not tested, but improves other tests :)

[Copilot]

Pull Request Overview

This PR improves error reporting in run_rllib_example_script_experiment by guarding against an IndexError when extracting nested exception messages and falling back to repr(e).

• Adds a safe extraction of e.args[0].args[2] with a fallback to repr(e) if the tuple is too short.
• Fixes a spelling typo in a comment (criteris → criteria).
• Updates the raised RuntimeError to include the new errors list.

Comments suppressed due to low confidence (1)

rllib/utils/test_utils.py:1352

• Add a unit test for this fallback branch to ensure that when nested exception args are too short, the code uses the fallback behavior as intended.

        # Might cause a Index error if Tuple is not long enough in that case use str(e)

[github-actions]

This pull request has been automatically marked as stale because it has not had
any activity for 14 days. It will be closed in another 14 days if no further activity occurs.
Thank you for your contributions.

You can always ask for help on our discussion forum or Ray's public slack channel.

If you'd like to keep this open, just leave any comment, and the stale label will be removed.

[Daraan]

(ง •_•)ง stale

[kamil-kaczmarek]

LGTM!