EVENT DOSSIER
On September 4, 2026, Google Cloud announced a serious network failure in its us-central1-b region. The incident occurred when an engineer, during routine hardware maintenance, failed to follow procedures and sequentially removed the fiber connections of all devices in that region between 07:41 and 11:52 Pacific Time. This incident caused a 100% decrease in resource traffic in the affected area, rendering some virtual machines inaccessible and increasing the packet loss rate. Although Google’s data centers have multi-routing redundancy and physical separation mechanisms to prevent single-point failures, this accident still resulted in service interruptions. After the failure, Google shifted traffic to other healthy capacities within the region. Technical personnel then reinserted the fibers to restore the connection, ultimately returning traffic to normal and services to operation.
Google Cloud 于 9 月 1 日因一名工程师在例行硬件维护中意外拔除所有光纤，导致 us-central1-b 区域部分服务出现严重网络降级和资源隔离。该事件从太平洋时间 07:41 持续至 11:52，受影响区域内资源流量下降率峰值达 100%，虚拟机不可达且丢包率升高。尽管 Google 数据中心设计具备多路由设备冗余及物理分离机制以抵御单点或双点故障，但此次事故系因维护人员未按程序操作，在 13 分钟内顺序拔除该区域所有设备的光纤路径所致。故障发生后，Google 将流量切换至区域内其他健康容量，技术人员随后重新插拔光纤恢复链路，最终使流量恢复正常并返回服务。