Yahoo Scale, Platform Engineering, and Reliability

Building for Reliability: Durable Execution & Insights from Temporal's Report - Preeti Somal

Topic 3 of 11

Published Oct 6, 2025 · Clip 5:34–12:02 (6:28)

Preeti details her Yahoo years (2008-2013) running internal cloud services, including the Yamas monitoring service consuming billions of time-series data points daily. She stresses treating scale and reliability seriously in design, testing, and incident response, and building a culture of strong engineering ownership despite infrastructure work being invisible. She shares Yahoo platform-team lessons still relevant to modern platform engineering: work closely with platform consumers, pick a few use cases, set adoption goals, and create visibility—even posting posters in kitchens. She also discusses reliability patterns like fault tolerance, alerting, monitoring, thundering herd mitigation, and rate limiting, noting Temporal now handles these concerns for developers.