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

{
#26233 is harder because it involves cache configuration precedence, initialization behavior, and distributed spend-accounting correctness across deployment modes, requiring broader regression testing. #24516 is comparatively localized to defensive handling and validation in an observability callback.
}
https://github.com/berriai/litellm/issues/26233 4:1 https://github.com/berriai/litellm/issues/24516