UModel (Universal Observability Model) is a data modeling method that uses a graph model for observable data. It addresses the core challenges of data collection, organization, and usage in enterprise environments. UModel uses a graph structure of Nodes and Links to describe the IT world. Through standardized data modeling, UModel enables unified representation, storage decoupling, and intelligent analysis of observable data. As the data modeling foundation for the Alibaba Cloud observability system, UModel provides businesses with a universal 'interactive language' for observability. This language allows people, programs, and AI to understand and analyze observable data to build true full-stack observability.
Core concepts
Graph model basics
UModel uses the basic concepts of graph theory. It uses a directed graph of Nodes and Links to describe an IT system:
Node: The core of a Node is a Set, which is a dataset. A Set represents a collection of similar entities or data, such as an EntitySet, MetricSet, or LogSet. A Node also includes the storage class (Storage) for the dataset, such as SLS, Prometheus, or MySQL.
Link: Represents the relationships between Nodes, such as an EntitySetLink, DataLink, or StorageLink.
Field: Constrains and describes the properties of Sets and Links. A Field includes more than 20 configuration items, such as name, type, constraint rules, and analysis features.
Information science ontology
The design of UModel is derived from ontology in information science. It creates a formal knowledge representation by clearly defining concepts and their relationships within the observability realm. UModel maps traditional ontology concepts to specific technical implementations:
EntitySet: Defines a collection of similar concepts, such as the "apm.service", "k8s.pod", and "infrastructure.host" entity types.
Entity: A specific instance of an EntitySet, such as the "order-service" service instance or the "web-pod-001" pod instance.
Field: Describes the features of an Entity and a Set, such as the "service_name", "version", and "status" fields of a service.
EntitySetLink: Describes the relationships between EntitySets, such as "calls", "contains", and "runs_on" relationship types.
Constraints: Ensure data consistency and integrity by defining Field properties. These properties include data type, value range, and validation rules.
Data fusion concepts
UModel shifts the focus from "unified storage" to "data fusion":
Storage abstraction: Decouples data modeling from specific storage systems using Storage and StorageLink. This approach supports multiple storage options, such as SLS, Prometheus, and MySQL.
Semantic unification: Defines unified field semantics through Fields. The CommonSchema standardizes common field definitions to ensure semantic consistency across data sources.
Association analysis: Builds cross-domain entity relationships through EntitySetLink to enable intelligent data association.
Dynamic mapping: Supports field mapping and dynamic transforms to remain compatible with existing data formats.
Entity-relationship model
UModel represents the IT world as an entity-relationship model:
Entity modeling: Abstracts system components into EntitySets, such as services, databases, hosts, and containers.
Relationship definition: Defines relationships between entities through EntitySetLink, such as "calls", "contains", and "runs_on".
Data association: Associates observable data, such as logs, metrics, and traces, with the corresponding entities through DataLink.
Graph construction: Automatically generates entity topology graphs and data relationship graphs. This process supports graph queries and graph analysis.