In recent years IoT has developed very rapidly. IoT devices are used to monitor and control physical objects to transform the physical world into intelligent spaces with computing and communication capabilities. Compared to cloud computing, fog computing is used to support latency-sensitive applications at the edge of the network which allows client requests to be processed faster. This study aims to propose a monitoring framework for containerized black box microservices in a fog computing environment to evaluate CPU overhead, as well as to determine the operating status, service characteristics, and dependencies of each container.
This study proposes a monitoring framework to integrate computing resource usage and run-time information from service interactions using a black box approach that seeks to integrate service-level information and computing resource information into the same framework. The proposed framework is limited to observing information monitoring after the server receives a request. This study uses JMeter to simulate user actions, which send requests to the server, and this research assumes the user knows the IP address of the server. For container monitoring methods in fog computing, all are indirect monitoring methods.
The results of this study indicate that the proposed framework can provide operational data for visualization that can help system administrators evaluate the status of running containers using a black box approach. System administrators do not need to understand and modify target microservices to gather service characteristics from containerized microservices. Regarding future research, it is suggested to expand the exploration of modified system information, and that part of the container management tool code can be pre-tried so that the framework proposed in this study can provide real-time quantitative indexes for the load balancing algorithm to help optimize the load balancing algorithm.