A detailed post‑mortem explains how a sudden traffic surge, insufficient pod count, and a custom thread‑pool bottleneck caused Tomcat thread‑pool saturation, health‑check failures, and a zone‑wide outage, and offers concrete lessons on capacity planning, monitoring, and safe coding practices.
This article details DeWu's container SRE approach, covering SRE fundamentals, on‑call response, SLO/SLA design, change management, capacity planning, kernel‑parameter monitoring, security safeguards, and a real‑world incident analysis, providing actionable insights for building resilient cloud‑native services.
This article systematically details Didi's migration of over 30 Elasticsearch clusters, 3,500 nodes and 8 PB of data from version 2.3.3 to 6.6.1, covering background, problem analysis, multi‑version architecture redesign, capacity planning, tiered storage, FastIndex, query replay, upgrade pitfalls, and the resulting cost reduction and performance improvements.
