Temporal Knowledge Graph Construction for microservice-based digital power applications

Modern digital power systems increasingly adopt microservices to enhance flexibility and scalability. However, in cloud-native environments, dynamic interactions across the microservices, Kubernetes, and physical machine layers complicate system monitoring and fault diagnosis. Key challenges include real-time high-frequency data collection, semantic integration of heterogeneous data, and incremental updates to knowledge graphs. This paper proposes constructing a multi-dimensional temporal knowledge graph by collecting logs, metrics, and traces from power systems and employing semantic models to unify data entities and relationships. A real-time incremental update mechanism significantly improves the efficiency of online monitoring and fault diagnosis in large-scale power systems.