fix(tests): Fix flaky container and scientific notation tests

[shin-bot-litellm]

Test Fixes

1. test_router_acreate_container_without_model

Problem: The test mocked container_create_handler (sync) but router.acreate_container uses _is_async=True which calls async_container_create_handler (async). This caused the test to hit the real OpenAI API with Authorization: Bearer None.

Fix: Use AsyncMock on async_container_create_handler instead.

2. test_register_model_with_scientific_notation

Problem: The test used a static "unique" model name (test-scientific-notation-model-unique-12345). When tests run in parallel with -n 16 in CI, this can cause conflicts between test workers accessing the shared litellm.model_cost dictionary.

Fix: Use uuid.uuid4() to generate truly unique model names per test run.

---

Fixes for CI stability

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
litellm	Ready	Preview, Comment	Feb 7, 2026 5:48pm

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

Shin seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[greptile-apps]

Greptile Overview

Greptile Summary

Fixed two flaky tests that were causing CI instability when running with parallel test workers.

• test_router_acreate_container_without_model: Corrected mock from sync container_create_handler to async async_container_create_handler using AsyncMock, preventing the test from hitting the real OpenAI API with invalid credentials
• test_register_model_with_scientific_notation: Replaced static model name with uuid-based unique names to prevent race conditions when tests run in parallel with -n 16
• Removed unused imports (_cached_get_model_info_helper, get_model_info) that are no longer needed after cleanup

Both fixes are targeted and correct, addressing the root causes of test flakiness without introducing new issues.

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• Both changes correctly fix test flakiness issues with targeted, well-documented fixes. The async mock change properly aligns with the code path execution, and the uuid approach definitively solves parallel test conflicts. No logic changes to production code.
• No files require special attention

Important Files Changed

Filename	Overview
tests/test_litellm/containers/test_container_api.py	Fixed async mock to match router.acreate_container's async call path
tests/test_litellm/test_utils.py	Used uuid for unique model names to prevent parallel test conflicts

Sequence Diagram

sequenceDiagram
    participant Test as test_router_acreate_container
    participant Router as router.acreate_container
    participant CreateContainer as create_container
    participant Handler as base_llm_http_handler
    participant AsyncHandler as async_container_create_handler
    
    Test->>Router: await router.acreate_container(name="Test Container")
    Router->>CreateContainer: create_container(..., async_call=True)
    Note over CreateContainer: Sets _is_async=True from kwargs
    CreateContainer->>Handler: container_create_handler(..., _is_async=True)
    Note over Handler: Checks _is_async flag
    Handler->>AsyncHandler: async_container_create_handler(...)
    Note over Test: Mock intercepts here (AsyncMock)
    AsyncHandler-->>Handler: ContainerObject
    Handler-->>CreateContainer: ContainerObject
    CreateContainer-->>Router: ContainerObject
    Router-->>Test: ContainerObject

Loading

[greptile-apps]

_{2 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}