Overload Protection Strategies in WeChat's Large‑Scale Microservices

WeChat safeguards its billion‑user microservice platform by detecting overload when average queue wait exceeds 20 ms and applying a two‑dimensional priority system—business and hourly‑hashed user priorities—adjusted via histogram‑based admission controls and upstream back‑pressure, ensuring stable service during massive traffic spikes.

Backend EngineeringMicroservicesPriority Scheduling

0 likes · 12 min read

Overload Protection Strategies in WeChat's Large‑Scale Microservices

ITPUB

Aug 18, 2022 · Operations

How WeChat Keeps Billions of Requests Stable: Overload Control Strategies for Massive Microservices

This article breaks down WeChat’s 2018 overload control system for massive microservices, explaining the problem of service overload, detection via average waiting time, and a multi‑level priority‑based mitigation strategy that dynamically adjusts admission thresholds to keep billions of daily requests stable.

MicroservicesPriority SchedulingWeChat

0 likes · 12 min read

How WeChat Keeps Billions of Requests Stable: Overload Control Strategies for Massive Microservices

WeChat Backend Team

Jan 17, 2019 · Operations

How We Built DAGOR: A Scalable Overload Control System for Massive Microservices

This article presents DAGOR, a decentralized overload control framework designed for large‑scale microservice architectures like WeChat’s backend, detailing its service‑agnostic design, priority‑based admission policies, adaptive algorithms, and experimental evaluation that demonstrates improved success rates, fairness, and robustness under heavy load.

MicroservicesScalable Systemsload shedding

0 likes · 49 min read

How We Built DAGOR: A Scalable Overload Control System for Massive Microservices