Year: 2024
Type: conference-paper
Abstract: The growing use of IoT and edge computing has highlighted the need for more efficient monitoring and visualization.These distributed systems have significant potential for lightweight, realtime monitoring, but face critical challenges, including inconsistent logging and performance tracking.Addressing these gaps is essential to optimize energy use and improve system performance in IoT-driven infrastructures.The objective of this paper is to present a solution in the form of an intelligent monitoring framework that is capable of recognizing and visualizing key operational scenarios in serverless edge devices within a clustered environment.By leveraging system logs and performance metrics, our approach provides a smart dashboard that aggregates, analyzes, and visualizes data in real-time, offering insights into system behavior.Our implementation demonstrates the ability to detect various scenarios using a Kubernetes-based cluster management system and data collection from Loki and Prometheus, along with visualization from Grafana.These capabilities include normal operation, overload, error handling, and load balancing scenarios.
Authors: Casper Bruun Christensen, Emily Treadwell Pedersen, Matthias Munch Jakobsen, Rasmus Louie Jensen, Sergio Moreschini, Junior Dongo, Davide Taibi, Michele Albano
Institutions: Aalborg University, Tampere University, University of Oulu
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Funder: Business Finland
Award: G36806144
OpenAlex ID: W4409242884