Meituan Origin Data Governance Platform: Architecture and Practices

Meituan, a highly digitalized and technology‑driven company, places great emphasis on extracting value from data. Over the years, its Meituan Hotel (酒旅) business has built a comprehensive solution consisting of a data warehouse and multiple data platforms (self‑service reporting, professional analytics, CRM, performance assessment, etc.) to eliminate data silos and support diverse analytical needs.

The early architecture (Figure 1) efficiently satisfied business demands but gradually revealed several consistency problems: inconsistent metric definitions, calculation logic, and data sources across platforms, leading to low trust in indicator data and poor decision‑making.

To address these issues, Meituan launched a data‑governance project and built the Origin Data Governance Platform, which records business processes, maps them to data processing and extraction, and enforces unified management of metrics and dimensions.

Challenges included determining where the platform should intervene in the existing stack with minimal intrusion, designing a concise yet effective management workflow, integrating various storage engines into a high‑concurrency, high‑availability data outlet, and ensuring data security across business lines.

Solution Approach positioned the governance layer between the data‑warehouse (or data‑mart) layer and the data‑application layer, acting as a bridge that provides rules, queryability, and monitoring. The new architecture (Figure 2) guides warehouse modeling, supplies metadata to applications, and reduces system intrusion.

Platform Architecture (Figure 3) consists of modular components: data storage, query, cache, metadata management, business management, security management, application management, and external APIs. Each module is designed for single responsibility, clear hierarchy, and maintainability.

Data Storage covers Hive, MySQL, Kylin, Palo, Elasticsearch, Druid, etc., with automated monitoring of table metadata changes, production status, and usage metrics.

Metadata Management handles both business metadata (metric definitions, dimension definitions) and data metadata (table schemas, model‑field bindings). It includes four sub‑modules: table management, model management, metric management, and dimension management.

Business Management is divided into business‑line management, theme management, and ticket (workflow) management, ensuring proper permission control, resource isolation, and traceability of data‑related requests.

Security Management integrates with Meituan’s internal permission system, providing page‑level access control, business‑line and data‑line user permissions, and API authentication.

Application Management comprises data applications, external applications, and a data map, enabling end‑to‑end tracking of data lineage from tables and models to downstream services.

External APIs expose metadata, data query, and monitoring/statistics interfaces, supporting multi‑SQL aggregation, cross‑engine queries, and performance monitoring.

The platform’s internal workflow (Figure 11) dynamically selects optimal models to generate SQL or query statements based on requested metrics and dimensions, leveraging a distributed query engine built on Akka Cluster, Redis‑backed task queues, and automatic load balancing.

Management processes (Figure 12) involve business owners defining metric business information, data engineers building tables and models, and finally delivering data services to users.

Since its launch, the Origin Data Governance Platform has supported more than ten internal data platforms, achieving unified metric and dimension management, a single data export point, unified monitoring and alerting, flexible query capabilities, and comprehensive data‑lineage visualization.

Future work will integrate the platform into Meituan’s broader “Tian‑Gong” ecosystem, standardizing metadata, query, and visualization interfaces to enable plug‑and‑play modules across the organization.