Incident Library?

API · MCP 한국어

Incident Library?

API · MCP 한국어

Component ?

NIC, cable and switch port link

NIC·케이블·스위치 포트 링크 · network-link

1 case · 2024

Root cause ?DNS or network fault

Contributing factor ?Unreviewed defaults

Symptom ?Link flapping NCCL timeout Throughput degradation

Cases, by severity

IL-0128 Network interfaces that dropped for a few seconds made NCCL time out and stalled or slowed LLM training on a cluster of more than 10,000 GPUsByteDance · 2024
Root cause DNS or network fault

View on the graph