Automatic service topology discovery
UModel discovers service-to-service call relationships and builds topologies:
-
Call relationship discovery: Identifies call relationships between services based on tracing data.
-
Dependency mapping: Creates a graph of service dependencies on external resources, such as databases and middleware.
-
Dynamic topology updates: Updates the service topology in real time to reflect changes in the system architecture.
-
Multilayer topology: Supports topology views at different granularities, such as service, instance, and infrastructure levels.
Cross-domain data association analysis
UModel correlates data across observation domains:
-
Application and infrastructure association: Correlates abnormal application performance with the status of underlying resources.
-
Log and metric association: Correlates log errors with abnormal performance metrics.
-
Security and business association: Analyzes the impact of security events on business.
-
Cost and performance association: Correlates resource costs with performance.
Full-stack observability and monitoring
UModel provides full-stack monitoring from frontend to infrastructure:
-
Application layer monitoring: Monitors and analyzes services, interfaces, and business flows.
-
Middleware monitoring: Monitors the performance of middleware, such as databases, caches, and message queues.
-
Infrastructure monitoring: Monitors underlying resources, such as hosts, containers, networks, and storage.
-
Business monitoring: Monitors business-level data, such as business metrics, user experience, and revenue impact.
Intelligent fault root cause analysis
UModel performs intelligent fault analysis based on an entity relationship diagram:
-
Multi-dimensional association analysis: Combines abnormal metrics, log errors, and trace anomalies for comprehensive analysis.
-
Spatio-temporal association analysis: Traces fault propagation paths across time and space dimensions.
-
Change impact analysis: Correlates change events, such as code releases and configuration changes, with faults.
-
Historical pattern recognition: Matches current faults against historical patterns for similarity analysis and prediction.
AIOps intelligent operations
UModel provides a standardized data foundation for AIOps:
-
Multi-dimensional association analysis: Correlates abnormal metrics, log errors, and trace anomalies for comprehensive analysis.
-
Anomaly detection: Detects anomalies across multiple dimensions using a unified data model.
-
Intelligent alerting: Reduces alert noise and improves alert quality using entity relationships.
-
Intelligent inspection: Regularly checks the health status of key entities and their relationships.