Modern web-based distributed systems are becoming increasingly complex and scaled up, making real-time anomaly detection more challenging. Coordination between components, semantic interpretation and adaptability are often lacking in traditional monitoring approaches. This paper introduces a Distributed Agent-based Intelligent Monitoring System (DAIMS) to address these challenges. It combines machine learning (ML) with semantic reasoning based on ontologies. The intelligent agents monitor service metrics locally and detect anomalies using XGBoost models. By collaborating through a shared knowledge base and message bus, they can infer system-wide problems. SWRL rules enable contextual inference as well as anomaly propagation analysis. Simulations were conducted under diverse workloads and fault injections. Compared to three widely adopted monitoring methods, DAIMS achieved a higher F1-score (0.91), a lower false positive rate (5%) and a higher detection rate (94%), while maintaining low inference latency (85 ms) and moderate overhead. Statistical tests were conducted to confirm the significance of these improvements. DAIMS achieves superior detection accuracy and responsiveness while requiring acceptable computational overhead. It was demonstrated that ML and semantic reasoning enhance fault explainability and robustness in distributed web environments.