Sorting LiteLLM GitHub issues by difficulty to implement. Higher = harder.

{
#33955 requires cross-cutting changes to batch processing, buffering, concurrency, resource limits, and regression coverage, with significant production-risk around large payloads and worker coordination. #28989 is comparatively localized to response translation and error propagation in two provider-specific paths.
}
https://github.com/berriai/litellm/issues/33955 5:1 https://github.com/berriai/litellm/issues/28989