EdgeInsight: An Agent-Driven Framework for System Health Monitoring in Edge Computing Platforms

This study introduces EdgeInsight, a robust system designed for effective performance monitoring in Edge computing environments. The system architecture is illustrated, based on an Edge Computing platform with System on Chip (SoC) technology. The SoC is equipped with integrated performance counters, enabling comprehensive monitoring and collection of performance parameters during computational workload execution. Computational workloads, tailored to specific use cases, operate within compute containers to address challenges related to library dependencies and facilitate seamless installation and migration. The core component of EdgeInsight, the lightweight agent, is deployed to track and monitor hardware and software events of computational workload in execution. Despite limitations imposed by SoC characteristics, the agent traces hardware and operating system events. Modern SoCs, featuring built-in sensors and performance counters, offer insights into internal system status from a power-and-performance perspective, covering energy consumption, temperature, and CPU components efficiency. With modern SoCs accommodating multiple cores on a single socket, EdgeInsight emerges as a versatile tool for modeling performance, energy usage, and temperature behavior in the dynamic realm of Edge computing.