Integração de ferramentas para a coleta de métricas em servidores Linux

Abstract

This work aims to find a solution for the instrumentation (the act of collecting and monitoring metrics) of Linux servers in micro and small organizations, where existing solutions in the market may not be a good option due to financial limitations and/or infrastructure. To achieve this goal, a study of the architecture of existing monitoring tools was done, in order to identify its main characteristics and functionalities. From this study, four basic functionalities were identified: collection, storage, visualization and alert triggering. For the selection of the tools, a research of metrics in computational systems and selection criteria was made, with the objective of selecting three softwares to be integrated. The selected softwares were Collectd (metrics collection daemon), InfluxDB (time series database) and Grafana (analytical visualization tool and alert triggering). These three softwares were integrated and the solution was validated in a real context of use, where it was possible to verify that all the initially planned functionalities were developed and had a satisfactory performance. During the validation, it was possible to make use of the tool for the collection, visualization and alerts of server metrics, increasing the efficiency of resource utilization during deploy of computational environments and being useful to identify problems, both in infrastructure and applications (sic).