[Bug fix] Router - handle cooldown_time = 0 for deployments

[ishaan-jaff]

[Bug fix] Router - handle cooldown_time = 0 for deployments

If a deployments cooldown is set to 0.00 or 0 then don't run cooldown logic for that specific deployment

Relevant issues

Pre-Submission checklist

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

• I have Added testing in the tests/litellm/ directory, Adding at least 1 test is a hard requirement - see details
• I have added a screenshot of my new test passing locally
• My PR passes all unit tests on make test-unit
• My PR's scope is as isolated as possible, it only solves 1 specific problem

Type

🐛 Bug Fix
✅ Test

Changes

[vercel]

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

Name	Status	Preview	Comments	Updated (UTC)
litellm	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jun 28, 2025 0:25am

[Copilot]

Pull Request Overview

A bug fix to ensure per-deployment cooldown_time = 0 bypasses default cooldown behavior and dynamic values are correctly propagated.

• Add time_to_cooldown parameter and early exit when it’s zero
• Propagate dynamic cooldown times through handlers and cache
• Update tests to cover zero, near-zero, None, and positive cooldown logic

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
tests/local_testing/test_router_cooldowns.py	Added tests to verify zero, near-zero, None, and positive cooldown behavior
litellm/router_utils/cooldown_handlers.py	Introduced time_to_cooldown, zero‐value early exit, and propagate to cache and callbacks
litellm/router_utils/cooldown_callbacks.py	Updated callback signature to accept Optional[float] cooldown_time
litellm/router_utils/cooldown_cache.py	Improved resolution between dynamic cooldown and default cooldown

Comments suppressed due to low confidence (3)

litellm/router_utils/cooldown_handlers.py:140

• This check is redundant because an earlier check already handles the case where deployment is None; consider removing the duplicate.

    if deployment is None:

litellm/router_utils/cooldown_handlers.py:101

• [nitpick] Consider renaming the parameter time_to_cooldown to cooldown_time to align with the terminology used elsewhere in the codebase.

    time_to_cooldown: Optional[float] = None,

tests/local_testing/test_router_cooldowns.py:164

• Add a test to verify that when a positive cooldown_time is provided in litellm_params, the cache is called with that value rather than the default.

    # Also verify the deployment is not in cooldown