In one of my previous #ABSQLNotes, I talked about computing latency via sys.dm_io_virtual_file_stats DMV – a critical metric to determine storage performance. Another way to measure latency is via the Average Disk Second Per Read perfmon counter. In simple words: how much time does it take to complete one read operation. And of course, this will be in milliseconds. The lower the better. The easiest way to compute this is to use the Performance Monitor counter: LogicalDisk\Avg. Disk sec/Read. There are many guidelines on the internet talking about preferred values for this counter. It ideally should be less than 10 ms. With today’s fast disks, even lesser. You need to create a baseline for this metric to get an average over time. And then slice it around for peak & non-peak hours. This value is computed for the entire disk and is not exclusive to SQL Server. But this is an important measurement of how fast the disks are able to complete their read operation. Sounds interesting? Watch the video below.

Get new content directly in your inbox – Click here

You can also visit DataPlatformGeeks (DPG) Video Library to watch more videos and to download video resources (presentation & demo code).

Hope the video was worth your time.

Happy Learning.

Learn more about SQLMaestros Products & Services: Video Courses| Hands-On-Labs | Learning Kits | Health Check Service