21CTO
Jul 30, 2016 · Operations
Building a 3‑Minute Fault Detection, 5‑Minute Recovery HA System for Games
This article explains how Alibaba’s NineGame platform achieved ultra‑high availability by shifting from system‑centric to business‑centric design, defining measurable goals (3‑minute issue detection, 5‑minute recovery, bi‑monthly incidents) and implementing a layered, automated, visual monitoring, client‑side retry, HTTP‑DNS, functional isolation, and multi‑site active‑active architecture.
Operationsbusiness‑centric designfault tolerance
0 likes · 22 min read
