Reliable Launches at Scale
summary
Google runs up to 180 production-impacting launches per week (7,000 total tracked), so a dedicated Launch Coordination Engineers team audits launches as auditors/gatekeepers, liaises across teams, drives technical blockers and transfers lessons learned. Reliability comes from a concise checklist automated into a self‑service tool (production advisor backed by a YAML spec), plus tactics that actually proved useful: feature flag frameworks, gradual rollout to 5–10% before full launch, and dark traffic to exercise backends safely.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 1:37What is a launch
- 4:14Launch tracking system
- 7:20Google Santa Tracker
- 8:51What makes a launch difficult
- 17:05How to make a launch work
- 20:01Launch Lead Engineer
- 21:38Launch Lead Engineer Roles
- 28:36How to Make Launches Reliable
- 29:16Checklists
- 30:56Production Advisor
- 32:16Architecture
- 34:24Capacity
- 36:06Reliability
- 36:56Route Monitoring
- 37:31Automation
- 37:54Performance Growth
- 38:14Dependencies
- 38:41Rollout
- 40:36Techniques
related talks
18mLeading High-Impact Engineering Launches
28mReliability Engineering: What we learnt so far
41mRelease management: Successful launches and updates on Google Play
28mFrom Startup to Scale Up: How to Scale Engineering teams
41mAn Introduction to SRE at Google (with DevOps Focus)
48mMoving Fast at Scale
42mHow DevOps Accelerates "Idea to Prod" at Google and What It Means to You
34mBringing Google Customer Reliability Engineering to Cloud Foundry