For query workloads, we can see important information about every single query that hits the system including details such as the total duration of the query, query text (MDX/DAX), start and end times, as well as the associated user account. We can also determine details as to how the query was executed such as the number of partitions scanned, aggregation hits/misses, cache hits/misses, other queries running at the same time, etc…all of which have an effect on the performance of any one particular query. A secondary benefit is that we’ll be able to identify the usage pattern(s) of folks using the cube. For example, is usage low/moderate throughout the week with a heavy spike on Friday mornings?
Bonus note: it looks like there will be an xEvents for Analysis Services GUI in SQL Server 2016.