A Multi-level Monitoring Framework for Containerized Self-Adaptive Early Warning Applications

Internet of Things (IoT) applications such as disaster early warning systems are increasingly being used in many areas of human life and business activities by using container-related technologies. However, there are still numerous technological challenges to be solved, particularly related to the time-critical Quality of Service (QoS) aspects of such applications. These types of systems should therefore be continuously monitored at various levels including infrastructure, container and application. Currently, there is a great lack of adequate monitoring systems for such a purpose. In this study, we present an architecture and implementation of our multi-level monitoring framework to ensure system health and adapt the performance of disaster early warning systems to runtime variations in running conditions. These changing conditions, which are out of service providers’ control, could be e.g. quantity, size and computational requirements of arrival requests to be processed or e.g. network quality (e.g. delay, packet loss and throughput) intrinsic to connections between individual application components deployed on different cloud infrastructures. Adaptation possibilities in this work include dynamically changing paths between containerized application components on one hand and also horizontal scaling of container-based application instances from the other side.