Development of a Task Monitoring System on a High Performance Computing System

The object of development is the task monitoring system on the Fisher supercomputer. The purpose of the development is to monitor tasks on the Fisher supercomputer using the installed monitoring tools. In the course of the work, a task monitoring system was developed, which allows collecting data on tasks running on a supercomputer, collecting average utilization rates of the components involved, and creating dashboards in the Grafana web interface for each task using the developed library. Plugins for the Telegraf background process were developed to collect the required metrics. The system interacts with the user through a web interface, where he can view the list of tasks, view detailed information about a specific task, and view graphs of the utilization of metrics involved in a specific task in the Grafana web interface. The resulting system will study the experimental architecture of Fisher supercomputer, will determine the efficiency of performing tasks and conveniently store data about the task with detailed visualization of metrics for further scientific research. The volume of the report of final qualifying work, not including the annex, is 41 pages. The number of illustrations in the work is 13, the number of sources of literature - 16.