Inside Meipai’s 3‑D Monitoring System: Scaling 150M Users with Unified Observability

Background

Meipai is a high‑profile mobile live‑streaming and short‑video community with more than 150 million monthly active users. The service runs on a hybrid cloud infrastructure that combines Meitu’s private cloud for core services with public‑cloud resources for disaster recovery and elasticity.

Monitoring Challenges

The scale and hybrid nature of Meipai’s platform create demanding monitoring and operations requirements. Over three years, the team needed to develop a monitoring solution that could provide end‑to‑end visibility, support rapid business growth, and maintain high reliability.

Three‑Dimensional Monitoring Architecture

The resulting system covers three layers:

Client‑side monitoring : metrics and logs collected from mobile apps to track performance, crashes, and user experience.

Server‑side monitoring : comprehensive instrumentation of microservices, databases, CDN, and big‑data pipelines, feeding into centralized metrics, tracing, and alerting platforms.

Unified reporting and dashboard : a single observability portal aggregates data from both private and public clouds, providing real‑time dashboards, anomaly detection, and historical analysis.

Key Technologies and Practices

The architecture leverages open‑source and commercial tools for metrics (Prometheus), tracing (Jaeger), log aggregation (ELK stack), and alerting (Alertmanager). Automation scripts ensure consistent metric collection across environments, while custom dashboards give product and operations teams actionable insights.

Speaker Profile

Wang Guansheng, former senior operations architect at Sina Weibo, now serves as Meitu’s Technical Director of Operations. He oversees DBA, service operations, big‑data operations, and CDN support, bringing extensive experience in scaling backend services from dozens to tens of thousands of machines.

Takeaways

The talk demonstrates how a large‑scale consumer app can build a holistic monitoring system that unifies client and server observability, supports hybrid cloud deployments, and enables proactive incident response, ultimately sustaining rapid product growth.