Urgent.News

What's breaking now, across thousands of outlets.

Tech

Google engineer unplugged every fiber they could see and – surprise! – took down a chunk of the G-Cloud

This sounds like a RTFM error at hyperscale

Google engineer unplugged every fiber they could see and – surprise! – took down a chunk of the G-Cloud

A Google Cloud engineer caused a partial outage by unplugging every fiber-optic cable they could see during routine hardware maintenance in the us-central1-b region on September 1. This action led to "severe network degradation and resource isolation," with traffic flow drop rates reaching 100% and virtual machines becoming unreachable.

Google's service health report explained that the company's datacenters are built to be resilient against single-device and most multi-device failures. However, if all cables are unplugged, redundancy is defeated. The immediate cause was an engineer's procedural error that unplugged 100% of fiber paths across devices within 13 minutes.

Once the engineer disconnected the routers, virtual machines in one zone lost contact with the outside world while still being able to communicate with each other. Google's resilience provisions automatically shifted traffic to healthy capacity elsewhere in the region once the issue was noticed. Technicians reseated the disconnected optical links, restoring traffic flow rates and normalizing the traffic.

The incident highlights the importance of following safety procedures when maintaining infrastructure, as a simple oversight like unplugging cables can have significant consequences.

Written by urgent.news from The Register Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at theregister.com →

More in Tech

Tenant-Aware Error Capture in NestJS: HTTP Filters, Cron Jobs, Queue Workers

An error-tracking setup for a NestJS backend has one hard problem, and it isn't which vendor you sign up with: HTTP exceptions arrive by the thousand, cron jobs and queue workers arrive by the…

  • Implement three-layer error capture system: global HTTP filter, cron job wrapper, queue worker catch
  • Append tenant ID and experiment cohort to error data at each entry point
  • Choose between Sentry, OpenTelemetry, or Infrai for consistent error tagging

개발자 문서 링크 저장을 위한 실용적인 체크리스트

개발자의 링크 보관은 검색보다 맥락에 가깝다. 프레임워크 문서, API 레퍼런스, GitHub 이슈, 변경 기록, 마이그레이션 안내, 패키지 설명, 다른 프로젝트의 예제까지 하루에도 여러 종류의 페이지가 작업 화면을 스친다. 필요한 순간에는 분명 유용했지만, 몇 주 뒤 같은 오류가 다시 나타나면 상황이 달라진다.

More from Friday 4 September →