Google Global Outage (2020)

2 min read Original article ↗

Context

By 2020, Google wasn’t just a search engine it was identity infrastructure for the internet. Gmail, YouTube, Drive, Meet, and countless third-party apps depended on Google’s authentication systems.

What Happened

On the morning of December 14, 2020, at approximately 6:55 AM PT, users around the world began seeing something unusual: “You are not signed in.” At first, it seemed like a glitch. Then Gmail stopped working. YouTube went down. Google Docs refused to load. Within minutes, it became clear Google’s authentication system had failed. Behind the scenes, a critical system responsible for managing user identity and login tokens had run out of storage quota. No storage meant no ability to verify users. No verification meant no access. And just like that, billions of users were effectively locked out of their digital lives. Source: https://status.cloud.google.com/incidents/6PM5mNd43NbMqjCZ5REh

Root Cause

An internal authentication system exceeded its storage quota, preventing it from issuing and validating login tokens.

Impact

Gmail, YouTube, Google Drive, Meet all went down globally Third-party apps relying on Google login also failed Millions of users locked out simultaneously Widespread disruption to work, education, and communication

Fix

Engineers identified the quota issue and restored capacity to the affected system. Within ~45 minutes, services began recovering globally.

Lessons Learned

  • Even internal quota limits can bring down global systems
  • Authentication is a single point of failure for ecosystems
  • Invisible backend systems can have massive user impact
  • Monitoring must include capacity exhaustion risks

Prevention

  • Implement automatic scaling for critical storage systems
  • Add alerts for quota thresholds well before exhaustion
  • Decouple authentication dependencies where possible
  • Design graceful degradation for login failures