To address the current challenges in network monitoring and management, such as poor real-time performance, high cost, low accuracy, and slow fault detection and recovery, a P4-driven network telemetry and resource scheduling platform is designed and presented in this paper. This platform embeds telemetry information within data packets, achieving real-time monitoring and anomaly detection of network states. It significantly enhances the granularity and timeliness of network monitoring and employs a path rerouting mechanism for rapid network recovery. The platform boasts high scalability and customizability, allowing on-demand adjustment of telemetry strategies and parameters. It provides low-overhead, business-level telemetry information, offering comprehensive support for diverse services and advancing network management and service quality. This work provides a technical foundation for constructing the next generation of intelligent, efficient, and reliable networks through continuous research and practical application.