test: repair two CI tests broken by intentional changes

[yuneng-berri]

TLDR

Problem this solves:

• litellm_router_unit_testing and e2e_ui_testing are red on staging since this afternoon
• feat(router): meter auto-router tier and prompt customization against the auto_router license feature #39674 renamed a router method the static-scan allowlist still named
• 9ba6cab (LIT-4738) paginated the Tags and Model Hub tables at 25 rows
• The scrolling spec seeds 40 rows and expected all of them on one page

How it solves it:

• Rename the allowlist entry to auto_router_capability_violation
• Select 50 rows per page before counting, as the Logs case already does

User Flow

No user-facing change. Both edits are to tests only

Relevant issues

Linear ticket

Pre-Submission checklist

Please complete all items before asking a LiteLLM maintainer to review your PR

• I have added meaningful tests
• The handful of test files covering my change pass locally, e.g. uv run pytest tests/test_litellm/<your_test_file>.py -v. Leave the suites (make test-unit-*, make test-unit) to CI: it finishes in ~15 minutes where a laptop takes an hour or more
• My PR passes all required CI/CD checks (e.g., lint, schema.d.ts sync check, etc.)
• My PR's scope is as isolated as possible; it only solves 1 specific problem
• I have received a Greptile Confidence Score of at least 4/5 before requesting a maintainer review (Greptile reviews automatically once the PR is opened; only comment @greptileai to re-request a review after pushing changes)

Screenshots / Proof of Fix

Before (9878436)

test_no_linear_scans_in_router

1. CircleCI litellm_router_unit_testing on the last staging-bound run: https://app.circleci.com/pipelines/gh/BerriAI/litellm/89202/workflows/2e9c2960-9fd4-4cba-bf73-5acb3101ca4f
2. Observed: Found O(n) linear scan pattern in router.py: auto_router_capability_violation() at line 8815. The allowlist still names the pre-rename heuristic_v2_router_limit_violation

Tags and Model Hub scrolling specs

1. CircleCI e2e_ui_testing on the same workflow
2. Observed: both Tags: no row paints past the box it lives in and Model Hub: no row paints past the box it lives in fail with Expected: >= 40, Received: 25 after three attempts. The tables now page at 25

After (da5af0c)

test_no_linear_scans_in_router

1. python -m pytest tests/router_unit_tests/test_router_index_management.py -k test_no_linear_scans_in_router -q
2. Observed: 1 passed, 8 deselected in 1.03s
3. CircleCI litellm_router_unit_testing on this PR reports the same test as success (5.98s)

Tags and Model Hub scrolling specs

1. Brought the UI e2e stack up with E2E_KEEP_ALIVE=1 ./run_e2e.sh from tests/e2e/ui (seeded postgres, mock LLM on 8090, proxy on 4000 serving the UI built from this branch)
2. E2E_UI_BASE_URL=http://localhost:4000 LITELLM_MASTER_KEY=sk-1234 PROXY_LOGOUT_URL=https://www.example.com MOCK_LLM_PORT=8090 npx playwright test --config playwright.config.ts tests/tables/tableScrolling.spec.ts -g "Tags|Model Hub"
3. Observed: 2 passed (28.2s)
4. CircleCI e2e_ui_testing on this PR: 139 passed (9.6m), and the whole build_and_test workflow succeeded: https://app.circleci.com/pipelines/github/BerriAI/litellm?branch=litellm_%2Fcircleci-pipeline-failure-afe848

Type

✅ Test

Caveats (if any)

QA runbook

• tests/e2e/ui/tests/tables/tableScrolling.spec.ts › Tags: no row paints past the box it lives in - with 40 tags and the page size raised to 50, every row stays inside the table's own scroll box

  • Create 40 tags: for i in $(seq 1 40); do curl -s -X POST http://localhost:4000/tag/new -H "Authorization: Bearer sk-1234" -H "Content-Type: application/json" -d "{\"name\": \"scroll-tag-$i\"}"; done
  • Open http://localhost:4000/ui/?page=tag-management, set the rows-per-page selector in the table footer to 50
  • Expect at least 40 rows in the table, none of them painting below the bottom edge of the table body or of the page
  • Sanity check: this test makes sense to add and is not hand-wavey (e.g., assert actual expected spend instead of just spend > 0) or potentially flaky

• tests/e2e/ui/tests/tables/tableScrolling.spec.ts › Model Hub: no row paints past the box it lives in - same check on the Model Hub table with 40 models

  • Create 40 models: for i in $(seq 1 40); do curl -s -X POST http://localhost:4000/model/new -H "Authorization: Bearer sk-1234" -H "Content-Type: application/json" -d "{\"model_name\": \"scroll-model-$i\", \"litellm_params\": {\"model\": \"openai/fake-gpt-4\", \"api_base\": \"http://127.0.0.1:8090/v1\", \"api_key\": \"fake-key\"}}"; done
  • Open http://localhost:4000/ui/?page=model-hub-table, set the rows-per-page selector to 50
  • Expect at least 40 rows, none painting past the table body or the page
  • Sanity check: this test makes sense to add and is not hand-wavey (e.g., assert actual expected spend instead of just spend > 0) or potentially flaky

Final Attestation

• The tests check the right things, including the edge cases, and regressions in the respective real-world customer use-cases are not possible after this PR

[greptile-apps]

Greptile Summary

This PR repairs two test suites after intentional production changes

• Updates the router linear-scan allowlist to reference the renamed capability-check method
• Selects a 50-row page before validating the 40 seeded rows in the Tags and Model Hub scrolling tests

Confidence Score: 5/5

The PR appears safe to merge because both changes preserve the tests' intended coverage after the underlying behavior changed

The pagination helper targets controls shared by both affected tables, and the router allowlist now matches the renamed method while preserving the previous exception

Important Files Changed

Filename	Overview
tests/e2e/ui/tests/tables/tableScrolling.spec.ts	Selects a supported 50-row page so both scrolling tests render and inspect all 40 seeded records
tests/router_unit_tests/test_router_index_management.py	Updates the existing linear-scan exception to match the renamed router method without broadening the allowlist

_{Reviews (1): Last reviewed commit: "test: repair two CI tests broken by inte..." | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!