litellm · Issues· 5170 open

[Bug]: Synchronous Azure deployment pricing can be recorded as $0

[Bug]: Streaming guardrails: value split across two SSE chunks can pass per-chunk checks

[Bug]: enforce_model_rate_limits — deployment TPM is never enforced for streaming requests (hidden_params.litellm_model_name is None)

[Bug]: MCP sending x-litellm-api-key to a dcr_bridge server returns an empty tool list

[Feature]: Support adding skills to private repos with authentication (SSH key for local, GitHub access token for remote/private)

[Bug]: Datadog LLM Obs embedding spans lack top-level model_name, causing "Partial cost"

[Feature]: Presidio output_parse_pii — derive placeholders from the value so tokens stay stable across requests

/v1/models does not respect Team member allowed_models restrictions

[Bug]: Issue using litellm with pi.dev

[Bug]: OTel v2: gen_ai.output.messages and gen_ai.response.finish_reasons missing on Responses API (/v1/responses) spans

[Bug]: /vertex_ai/live cost/logging doesn't resolve model_group_alias

[Bug]: Pause (blocked) and delete are ignored at routing time for wildcard-model deployments — PatternMatchRouter pool is never reconciled

[Feature]: unlimited SSO Login with Standard Plan

adaptive_router: thompson_sample has no alpha/beta floor — one persisted alpha=0 cell bricks the router with HTTP 500 (gammavariate)

[Bug]: Gemini native tool history needs canonical functionCall and functionResponse keys