### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

A bug happened! I attempted to use 'A/B Testing - Traffic Mirroring', as documented here: https://docs.litellm.ai/docs/traffic_mirroring. The call to the primary model was successful, but the call to the secondary model appears to have never happened and fails silently. I tested this with multiple model combinations from OpenAI and Anthropic. The results were the same. No log entries are found for the secondary/mirrored model call. Direct calls to the secondary/mirrored model succeed as expected.

Also, note that the docs state: "Use the `is_silent_experiment: True` flag in your logs to filter and compare results between the primary and mirrored calls." Using `DEBUG` level logging, the term `is_silent_experiment` never appears.

My LiteLLM Proxy configuration (snippet):

```ini - model_name: bedrock-mantle-openai-gpt-5-5 litellm_params: model: bedrock_mantle/openai.gpt-5.5 api_key: os.environ/AWS_BEARER_TOKEN_BEDROCK silent_model: openai-gpt-5-5

- model_name: openai-gpt-5-5 litellm_params: model: openai/gpt-5.5 api_key…