test(key-management): unwind the global writes the key tests scaffold around

[yuneng-berri]

TLDR

Problem this solves:

• 17 tests hand-roll save, try, write, restore around a global
• 4 more restore a flag to a hard-coded value, not the old one
• Every one of them is a TQ005 violation

How it solves it:

• monkeypatch.setattr, so the scaffolding comes out
• The fixtures stop guessing what they restore to

User Flow

No end-user behavior changes. A proxy admin calling
https://litellm-domain/key/generate gets the same limits enforced the same way;
this is the test file around that endpoint, not the endpoint.

Relevant issues

Linear ticket

Pre-Submission checklist

• I have added meaningful tests
• The handful of test files covering my change pass locally
• 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

Screenshots / Proof of Fix

Shared setup: run the file, then read the three globals back in the same
interpreter, and count what the quality rules see.

uv run python -c '
import litellm, pytest
names = ("upperbound_key_generate_params", "default_key_generate_params", "enable_key_alias_format_validation")
before = {n: getattr(litellm, n) for n in names}
pytest.main(["tests/test_litellm/proxy/management_endpoints/test_key_management_endpoints.py", "-q", "-p", "no:randomly"])
for n in names:
    print(f"litellm.{n} after the run:", "restored" if getattr(litellm, n) == before[n] else "LEAKED")
'

Before (ff02d5c)

1. The command above

443 passed, 4 warnings in 2.66s
litellm.upperbound_key_generate_params after the run: restored
litellm.default_key_generate_params after the run: restored
litellm.enable_key_alias_format_validation after the run: restored

2. uv run python scripts/check_test_quality.py tests/test_litellm/proxy/management_endpoints/test_key_management_endpoints.py | grep -o 'TQ00[0-9]' | sort | uniq -c

  24 TQ001
  16 TQ002
   1 TQ003
  40 TQ005

After (1f682f8)

1. The same command

443 passed, 4 warnings in 2.29s
litellm.upperbound_key_generate_params after the run: restored
litellm.default_key_generate_params after the run: restored
litellm.enable_key_alias_format_validation after the run: restored

2. The same count

  24 TQ001
  16 TQ002

3. uv run python scripts/test_quality_gate.py

OK: every TQ rule is within its test-suite ceiling (base origin/litellm_internal_staging)

Type

🧹 Refactoring

✅ Test

Caveats (if any)

• Both sides read "restored", so no live leak is being fixed here
• The conftest snapshot was already catching all three globals
• The point is the tests stop depending on it, which WS1 wants
• TQ001 and TQ002 in this file are untouched and need real work
• Touches test-quality-budget.json, as do test: drop the cwd-relative sys.path.insert calls from the test suite #37802, test: unwind environment writes in tests/test_litellm with monkeypatch #37806 and test(cost-calc): stop 182 global writes leaking out of the cost-calc suites #37815

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 test-only refactor replaces manual writes and restoration of key-management globals with pytest monkeypatching and removes an ineffective path mutation

• Restores key-generation defaults, upper bounds, and alias-validation flags to their actual prior values after each test
• Lowers the TQ003 and TQ005 quality budgets to reflect removed violations
• Leaves production key-management behavior and test assertions unchanged

Confidence Score: 5/5

The PR appears safe to merge because it only improves test isolation and ratchets matching quality budgets

The changed tests preserve their configured values and assertions while pytest now restores global state even when a test exits exceptionally

Important Files Changed

Filename	Overview
tests/test_litellm/proxy/management_endpoints/test_key_management_endpoints.py	Replaces manual global mutation scaffolding with function-scoped monkeypatch restoration without weakening the existing assertions
test-quality-budget.json	Ratchets TQ003 and TQ005 ceilings to the reported post-refactor counts

_{Reviews (1): Last reviewed commit: "test(key-management): unwind the global ..." | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!