eBPF-Based Network Monitoring Platform on Kubernetes

As enterprises transition server maintenance from local to cloud services, microservices architecture plays a crucial role. Therefore, we propose solutions to address key challenges in microservices architecture, including the efficient acquisition of customized monitoring metrics, load balancing based on the operational status of microservices, optimization of system resource utilization, and the rate limiting mechanisms to mitigate the impact of instantaneous traffic surges. In the paper an event-driven network monitoring platform based on extended Berkeley Packet Filter (eBPF) within the microservices architecture is developed to achieve these objectives. Furthermore, integrating these solutions with Kubernetes can enhance the scalability, reliability, and manageability of the microservices environment.