Introduction to X Monitoring System: Architecture, Modules, and Implementation Details

The monitoring system is described as a security guard for application systems, capable of monitoring both business services and server status, allowing early detection of issues to reduce business loss.

The author introduces the internally developed X Monitoring system, which consists of three main modules: Configuration, Reporting, and Monitoring, spanning job, API, and monitoring systems and utilizing middleware such as Redis and Qbus.

Because of the large volume of monitoring data and the need for long‑term historical dashboards, the system stores monitoring data in ElasticSearch and configuration information in MySQL.

An architecture diagram (image) illustrates the overall design.

X Monitoring provides two types of monitoring: server‑side (business) monitoring and agent‑side (server) monitoring.

Server‑side monitoring requires no deployment on target machines; it gathers status via HTTP requests and database connections, covering API monitoring (configurable endpoints with alert thresholds) and DB monitoring (SQL queries to generate reports and trigger alerts).

Agent‑side monitoring requires a lightweight PHP tool package on the target machine, minimizing dependencies and enabling cross‑language agents.

The agent follows two main processes: periodic restart of the monitoring service to sync configuration changes and prevent process death, and the monitoring workflow that collects data and reports it to the monitoring API.

Server‑side monitoring includes file monitoring, process heartbeat, CPU, memory, disk, and load metrics.

Each metric is described with its purpose and example commands (illustrated by images): • File monitoring checks for the existence of a specified file. • Process heartbeat verifies whether a given process is alive. • CPU monitoring records usage data for reporting. • Memory monitoring records free memory, buffers, and cache. • Disk monitoring gathers usage information. • Load monitoring captures average system load over 1, 5, and 10 minutes, with threshold guidance for healthy versus problematic states.

The article concludes that X Monitoring, while simple, provides core monitoring, reporting, and alerting functions that meet business needs, and encourages developers and testers to collect and implement required monitoring to ensure product stability.