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

{
The left issue requires tracing and correcting shared asynchronous routing, retry propagation, exception metadata, cooldown selection, and regression coverage across multiple provider-facing paths. The right issue is a localized dashboard rendering change with comparatively limited integration risk.
}
https://github.com/berriai/litellm/issues/35071 5:1 https://github.com/berriai/litellm/issues/30468