#33960 Per-group allowed_fails / cooldown_time / enable_health_check_routing via routing_groups
### Check for existing issues
- [x] I have searched the existing issues and checked that my issue is not a duplicate.
### The Feature
Extend RoutingGroup (**litellm/types/router.py**) with optional **allowed_fails**, **allowed_fails_policy**, **cooldown_time**, **enable_health_check_routing** fields, resolved via the **Router._model_to_group map** that **_init_routing_groups** already builds — falling back to the existing router-wide scalars when a group doesn't override them. I named the exact two functions in **cooldown_handlers.py** and the two filter methods in router.py that would need the group-lookup added, plus a sample config.
### Motivation, pitch
<html><head></head><body><h2>Background</h2><p>We run LiteLLM as a shared internal gateway in front of multiple services. While all services communicate through the same proxy, each service routes to a different model or model group depending on its use case:</p><ul><li><p><strong>Latency-sensitive inline scoring service</strong> → optimized for low latency and fast failover.</p></li><li><p><strong>Batch/analysis workloads</strong> → can tolerate transient failures and prioritize capacity.</p></li><li><p><strong>Chat as…