Key Info

LiteLLM now includes rate-limit attribution in its responses, telling callers which exact limit caused a 429 error or triggered a fallback. The change is covered by PRs #51228 and #51138.

Highlights

  • When no fallback is available, LiteLLM returns a 429 with rate_limited_by set to the exact rule that blocked the request (e.g., model_per_end_user).
  • When a fallback is used (e.g., Opus hits its limit and routes to Sol), response headers indicate that one fallback was attempted and which rate limit triggered it.
  • This makes debugging rate-limit issues and fallback behavior more transparent for gateway users.