Centralised monitoring and alerting solution for complex information management infrastructure

Abstract

Monitoring and alerting receive increasing attention over the last years together with growth of the information generated and managed from one side and increasing of the computing power and capacity of the computing hardware. Monitoring solutions have been tightly linked to the software being monitored making the broad view of the performance of the all interlinked services too complex and ineffective. In this work a monitoring and alerting solution based on “Prometheus” is developed. Continuous collection of metrics from various different services is performed and organized for visualization and monitoring in several layers of precision. Different exporters for various systems were evaluated and many of them enhanced. Alerting logic in response of detected performance problems and function irregularities has been developed and implemented with “Alert Manager”. Another software is developed for recording and visualizing current or past alerts and also is used as debugging tool of the alerts configuration. Visualization is implemented for Grafana with several dashboards. All tools and software packages used for implementing this monitoring and alerting solution are open source and free to use.