Andi Supply Chain, a subsidiary of the Midea Group, is a supply chain technology company dedicated to providing end-to-end digital supply chain solutions for manufacturing enterprises. Its services span warehousing, transportation, distribution, and last-mile delivery, serving as a core enabler of Midea’s digital transformation.

With rapid business expansion, Andi’s IT landscape has continued to scale:

• 69 core business systems supporting end-to-end operations

• Over 3,000 hosts (bare metal + virtual machines)

• Accelerating migration toward Kubernetes-based containerized microservices architecture

As system complexity and dynamism increase, the operations team faces several key challenges:

Fragmented Observability

Existing APM platforms lack sufficient integration across logs, traces, and metrics, resulting in scattered and loosely correlated data that prevents a unified observability view.

Inefficient Incident Resolution

During production incidents, engineers must switch between multiple tools and manually correlate information, significantly extending mean time to recovery (MTTR).

Monitoring Blind Spots in Container Environments

Hundreds of dynamically changing application instances make traditional agent-based monitoring and manual probe deployment unsustainable.

Alert Storms Without Intelligent Noise Reduction

Large volumes of raw alerts without AI-based deduplication or root cause identification make it difficult to surface truly critical issues.

In addition, applications operate as “black boxes,” lacking code-level visibility. Production anomalies are difficult to reproduce, requiring guesswork and waiting. Development, testing, and operations teams lack a unified collaboration platform, while traditional infrastructure monitoring cannot meet business-level observability needs.

Full-stack Probe Deployment & Data Collection

• APM deployment with 3,000 agent licenses covering all core business systems

• RUM coverage of 500,000 daily sessions across 69 systems and ~20 business departments

• eBPF-based non-intrusive network performance analysis and continuous process profiling

• Integration of existing Zabbix monitoring data to preserve legacy investments

Unified Observability Data Fusion

Bonree ONE establishes a unified data model integrating metrics, traces, and logs, automatically generating global service topology with health-state visualization. It also supports native PromQL queries and seamless migration of Prometheus alert rules.

End-to-End Digital Experience Monitoring

The platform reconstructs the full user journey from application entry to exit, analyzing crashes, latency, JavaScript errors, and network requests to precisely identify frontend experience bottlenecks.

Intelligent Alerting and Root Cause Analysis

Supports both AI dynamic baselines and static thresholds for anomaly detection, enabling intelligent alert suppression. A two-stage root cause analysis mechanism significantly reduces incident diagnosis time.

Non-Intrusive eBPF Deep Monitoring

Provides kernel-level continuous profiling, CPU hotspot detection, and L4/L7 network traffic analysis with topology visualization, filling monitoring gaps in containerized environments.

Visualization & System Integration

Includes five customized operational dashboards and integrates with Midea’s Meixin unified alert platform and LDAP authentication system.

Project Value and Benefits

Operations Value

Code-level performance bottleneck identification and AI-driven root cause analysis significantly reduce MTTR, shifting teams from reactive firefighting to proactive optimization.

Business Value

Enhanced observability and real-time business insights improve user experience and provide a stable foundation for business growth.

Development Value

Bonree ONE breaks data silos across development, testing, and operations teams, improving collaboration efficiency and accelerating delivery cycles.

Management Value

Custom operational dashboards make service quality metrics transparent and quantify IT contributions to business outcomes, supporting data-driven decision-making.

Going forward, Bonree Data will continue to deepen its collaboration with Andi Supply Chain to:

• Expand AI Workspace and intelligent agent applications

• Explore more automated operations decision-making scenarios

• Complete log module integration as the final piece of the unified observability stack

• Extend root cause analysis and AI Q&A capabilities to broader business scenarios, continuously enhancing intelligent operations efficiency

Why Bonree

Full-stack End-to-End Observability

Bonree ONE provides complete observability coverage from user experience, application services, middleware, databases, to underlying infrastructure.

Strong Technological Leadership

The company holds:

• 60 authorized invention patents

• 136 software copyrights

• 32 core proprietary technologies