First published on MSDN on Jul 08, 2008
The mistake I see people make is when looking at the SQL Server PAGE I/O waits and stalled I/O warnings is when comparing it to the ...
Updated Jan 15, 2019
Version 2.0
Blog Post
If you're analyzing SQL Server I/O bottlenecks, this is a crucial point that many people overlook. Looking at Avg. Disk Sec/Transfer might give the impression that everything is fine, but the real issue is with spikes that can mislead the overall average. Your example clearly shows how even a single high-latency I/O operation can have a major impact. That’s why analyzing data points like PAGE I/O waits and Virtual File Stats is essential.
By the way, if you're dealing with such bottlenecks, a https://pcbottleneckcalculator.org/ can be really useful to gain insights based on workload and latency.