Incident Library

API · MCP 한국어

API and MCP kubegap Corrections and proposals

Incident Library

API · MCP 한국어

Component

Model request routing

모델 요청 라우팅 · model-routing

3 cases · 2022 to 2025

Root cause Software defect3

Contributing factor Automation amplification1 Excessive blast radius1 Ignored early signal1 Limited debug access1 Monitoring blind spot1 No alert on limit1 No headroom1 Overlapping failures1

Symptom High error rate2 Service outage2 Probe failures1 Misrouted requests1

Organisation Anthropic1 Google Cloud1 OpenAI1

Cases, by severity

IL-0022 Three overlapping infrastructure bugs (context-window misrouting, TPU output corruption and an XLA top-k miscompilation) degraded Claude's responsesAnthropic · 2025
Root cause Software defect
IL-0123 A response buffer allocated anew in a loop made OpenAI's routing layer nodes hit their memory limits and fail readiness checks on a morning with more completions than any day before, causing 502 and 503 errors across all modelsOpenAI · 2023
Root cause Software defect
A faulty resource cleanup process marked Vertex AI prediction endpoints as deleted worldwide, undeploying customers' models and failing predictions for 4 hours 20 minutes

View on the graph

Resources deleted1

Response quality degradation1

IL-0132

Google Cloud · 2022

Root cause Software defect