How 360 Mobile Guard Optimizes App Performance with Its Own Monitoring System

Preface

With the rise of mobile internet, app performance has become a critical factor for rapid growth. 360 Mobile Guard, a super‑app with hundreds of plugins built on its own RePlugin framework, has constructed a complete performance monitoring system to address these challenges.

What performance issues do apps have?

App performance problems can be broadly categorized into several areas such as startup time, UI smoothness, memory usage, network latency, power consumption, and stability.

Mobile Guard Performance Control System

The system consists of three parts: offline, online, and supplementary methods.

Offline Scenario

1. Code Red Line

Includes code standards, Lint, FindBugs, and package size monitoring. 360 Mobile Guard enforces strict coding guidelines, uses custom Lint rules to detect over‑draw, and custom FindBugs rules to catch memory leaks.

2. Argus APM Debug Mode

A performance tool based on Argus APM for developers and testers. Features:

Real‑time performance data collection

Local analysis

Warning notifications

Traceable performance issues

Multi‑process display

Integration with performance bug tracking

When a problem is detected, logs are stored on the SD card for later analysis.

3. Testing

Rigorous testing includes functional and automated tests to catch bugs and performance issues such as power consumption and stability before release.

Online Performance Monitoring Scenario

1. Crash Backend

Collects crash data in real time, groups them by hash, and sends email alerts for new or surging crashes. Regularly addresses top‑crash items to reduce overall crash rate.

2. Argus APM Performance Monitoring Platform

A non‑intrusive, comprehensive monitoring platform that supports plugin monitoring, cloud‑based control, targeted APM activation, and real‑time data collection and analysis.

Key monitoring modules include:

App startup time (cold and hot)

Activity lifecycle duration

Network request performance

Memory usage across devices and Android versions

ANR detection

UI jank detection

Process information and app survival rate

File/database size checks

IO access statistics

Thread usage

Frame rate monitoring

Power consumption (especially on 360OS)

Sensor usage analysis

CPU usage statistics

The Argus APM data backend supports massive data queries, real‑time results, multi‑dimensional filtering, and instant alerts.

Other Supplementary Methods

1. Diagnostic Plugin

For online user feedback, a diagnostic plugin collects performance data under the guidance of support staff, enabling rapid problem localization and resolution.

2. User Feedback Backend

Aggregates user‑reported issues over time to reveal trends and uncover problems that automated monitors may miss.

Performance Case Studies

Automatic exit issue

Network traffic anomaly

Abnormal power consumption case 1

Abnormal power consumption case 2

Future Directions

1. Deep Learning Exploration

Integrate deep learning and big data techniques into performance analysis for smarter optimization.

2. Automation and Tooling

Adopt automated and tool‑driven approaches to make performance monitoring more convenient and efficient.

3. Open Source / Sharing

Plan to open‑source the Argus APM solution, contributing it to the open‑source community.