This study introduces a dynamic, ontology-driven approach to modeling Kubernetes clusters by leveraging an in-memory dependency graph. This graph adeptly captures live structural and behavioral links, such as pod-to-node allocations and service-to-pod mappings, directly sourced from the Kubernetes API. This model enables efficient detection of environmental shifts, ensuring that only relevant changes are relayed to the OWL-based ontology, thereby optimizing update performance. We enhance the ontology’s capabilities with OWL RL reasoning and tailored SWRL rules, allowing for the deduction of implicit information, detection of inconsistencies, and facilitation of rich SPARQL-based queries for semantic oversight. Through a range of controlled tests and simulated cluster scenarios, we demonstrate the model’s ability to uphold logical consistency, adjust to runtime dynamics, and maintain high inferential precision. The proposed system shows tangible improvements in update speed and query performance. Our results suggest that this integrated strategy successfully merges cloud-native operational data with semantic frameworks, advancing intelligent, ontology-based observability and decision-making in Kubernetes settings.