test(evaluator): Keep Evaluator mock VirtualModels out of orphan cleanup

[JashG]

Issue: Autoprovisioned VMs can be deleted mid-job by the models controller, causing flaky llm-judge ModelRef failures with VirtualModel 404s.

Fix: Creates test VirtualModels with autoprovisioned=False so they survive for the lifetime of the job.

Example test failure: https://github.com/NVIDIA-NeMo/nemo-platform/actions/runs/29863264284/job/88746335863

Summary by CodeRabbit

• Tests
  • Updated evaluator end-to-end test setup to use explicitly configured virtual models.
  • Improved test reliability by preventing test resources from being incorrectly treated as orphaned during evaluation.

[coderabbitai]

📝 Walkthrough

Walkthrough

Evaluator E2E setup now disables virtual-model autoprovisioning when creating mock providers and preparing ready models, including marking passthrough models as non-autoprovisioned.

Changes

Evaluator provisioning

Layer / File(s)	Summary
Disable virtual-model autoprovisioning e2e/test_evaluator_plugin.py	Mock provider creation and ready-model preparation explicitly disable virtual-model autoprovisioning and mark the passthrough model as not autoprovisioned.

Possibly related PRs

• NVIDIA-NeMo/nemo-platform#735: Refactors evaluator E2E mock and ready-model provisioning paths.

Suggested reviewers: briannewsom

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: keeping Evaluator mock VirtualModels out of orphan cleanup.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch jgulabrai/evaluator-e2e-flakiness

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Suite	Lines Covered	Line Rate	Branch Rate
Unit Tests	26399/34046	77.5%	61.8%
Integration Tests	15162/32671	46.4%	18.6%

[JashG]

Covered by #754 so closing