IT Operations Analytics: Root Cause Analysis via Complex Event Processing

IT operation analytics (ITOA) is used for discovering complex patterns in data from IT systems. The analytics process still includes a significant portion of human interaction which makes the analysis costly and error-prone. Human operators need to formulate queries over the collected data to identify the complex patterns. Since the queries describe complex relations, the queries are usually multilevel, perplexing, and complicated to create. For the querying the complex relations, complex event processing methods are successfully used in other domains. In this paper, we demonstrate an application of the complex event processing principles in the ITOA domain. We adjust T-Rex complex event processing engine and improve TESLA event processing language to suit for ITOA tasks. Our demonstration includes two real-world use-cases. We show the utilization of the complex event processing for root cause analysis and demonstrate the natural formulation of complex queries that results in the reduction of the volume of the required human interaction.