Topic hub pagination
Continue browsing this topic cluster with SEO-safe static pagination.
Fix LiteLLM wrong Vertex AI URL construction for us/eu multi-region endpoints Includes evidence for LiteLLM troubleshooting demand.
LiteLLM Vertex AI: wrong base URL for us/eu multi-region endpoints Fix LiteLLM GitHub Copilot auth loop when access token becomes stale Includes evidence for LiteLLM troubleshooting demand.
LiteLLM GitHub Copilot authenticator: stale access token causes unrecoverable auth loop Fix LiteLLM rate limit response leaking API key hash in error messages Includes evidence for LiteLLM troubleshooting demand.
LiteLLM rate limit 429 response leaks full SHA-256 token hash in error.message Fix LiteLLM breaking DeepSeek V4 Pro multi-turn conversations by stripping reasoning_content Includes evidence for LiteLLM troubleshooting demand.
LiteLLM BadRequestError: reasoning_content must be passed back to the API (DeepSeek V4 Pro multi-turn) Fix LiteLLM structured output not working with Anthropic models on Bedrock Converse API Includes evidence for LiteLLM troubleshooting demand.
LiteLLM Structured Output fails for Anthropic models on bedrock/converse: Extra inputs are not permitted Fix LiteLLM false BudgetExceededError caused by team key spend being attributed to personal user spend Includes evidence for LiteLLM troubleshooting demand.
LiteLLM BudgetExceededError: Current cost exceeds Max budget due to team key spend polluting personal spend Fix LiteLLM budget tracking double-counting usage from previous month in current month Includes evidence for LiteLLM troubleshooting demand.
Duplicate Usage Aggregation Across Billing Cycles — April usage data double-counted in May budget Fix Claude Code multi-turn conversation failing through LiteLLM proxy gateway Includes evidence for LiteLLM troubleshooting demand.
API Error: 400 litellm.BadRequestError: 178 validation errors on multi-turn conversation Fix LiteLLM not tracking cost and token usage for Vertex AI batch jobs Includes evidence for LiteLLM troubleshooting demand.
Vertex AI batch jobs always record spend=0, prompt_tokens=0, completion_tokens=0 Fix LiteLLM rate limit error exposing API key hash in response body Includes evidence for LiteLLM troubleshooting demand.
Rate limit exceeded for api_key: <sha256_hash>. Limit type: parallel_request_limit Fix LiteLLM Redis user_api_key_cache deserialization error for team-scoped virtual keys Includes evidence for LiteLLM troubleshooting demand.
CacheCodec.deserialize: validation failed for LiteLLM_UserTable (1 validation error: user_id Field required) Fix LiteLLM 429 error response leaking token hash in rate limit error message Includes evidence for LiteLLM troubleshooting demand.
LiteLLM rate limit error message body leaks full SHA-256 token hash on 429 responses Fix LiteLLM Redis user_api_key_cache deserialization failure for team-scoped keys Includes evidence for LiteLLM troubleshooting demand.
LiteLLM Redis user_api_key_cache deserialization always fails for team-scoped keys (LiteLLM_UserTable.user_id required) Fix LiteLLM asyncio.CancelledError when listing tools for MCP server with HTTP transport and OAuth2 Includes evidence for LiteLLM troubleshooting demand.
LiteLLM asyncio.CancelledError listing tools for MCP server with HTTP + OAuth2 Fix LiteLLM returning empty tools list when upstream MCP server returns 401 Includes evidence for LiteLLM troubleshooting demand.
LiteLLM returns 200 {tools:[]} for 401 upstream OAuth errors Fix LiteLLM proxy 401 authentication error when virtual API key expires or budget is exhausted Includes evidence for LiteLLM troubleshooting demand.
litellm.AuthenticationError: 401 — Proxy virtual key expired or budget exhausted Fix LiteLLM performance degradation after upgrading to v1.81.x Includes evidence for LiteLLM troubleshooting demand.
Significant performance regression after upgrading from 1.80.5 to 1.81.x (UI + API slowness) Fix LiteLLM proxy memory leak causing high RAM usage over extended operation Includes evidence for LiteLLM troubleshooting demand.
Heavy RAM Usage over time Fix LiteLLM bulk invite generated API keys that don't have sk- prefix and get rejected Includes evidence for LiteLLM troubleshooting demand.
Bulk invite API keys missing sk- prefix, rejected on any call Fix LiteLLM llm_requests_hanging alert spam when requests are not actually hanging Includes evidence for LiteLLM troubleshooting demand.
llm_requests_hanging alert fires continuously for requests below 600s threshold