Stairway to SQL Server Extended Events
Erin Stellato, a Principal Consultant with SQLskills.com, explores the use of Extended Events as a diagnostic data collection tool or SQL Server. She describes how to define efficient low-overhead event sessions that exploit fully the vast number of events, as well as the powerful filtering and data collection options, offered by this new event collection infrastructure. She also demonstrates simple techniques to analyze event data and identify and troubleshoot the causes of poor SQL Server performance, such as long-running queries that consume vast amounts of CPU and I/O resources. It is time to embrace Extended Events and understand all that it has to offer, and Erin's stairway is the perfect place to start.
Over the course of this stairway series, we're going to explore in detail the use of Extended Events as a diagnostic data collection tool, to track down causes of poor performance in SQL Server. This first level will start from a point known and familiar to many DBAs, namely the use of SQL Trace to track down and investigate long-running queries.
In this Level, we'll walk through the basics of using the New Session dialog in the UI to create a new event session, define its events, actions and predicates, and establish a target for the session in which to collect the event data.
In the next installment of our stairway to Extended Events, we delve into the Live Data and Target Data Viewers.
In this level, we're going to dig a little deeper into the Extended Events engine, its architecture, and fundamental components. It will give you a deeper understanding of why, in general, an Extended Events session is inherently lower in overhead than an equivalent SQL Trace. We'll also investigate how to design our event sessions to minimize any unnecessary overhead during event data collection, even when we need to create relatively complex event sessions to investigate difficult performance problems.