NEW: Rule of Thumb Series
Server Performance Monitoring: Looking at the Big Picture with MIC


Looking at the Big Picture of Windows Server Performance

What do you really need to monitor on your Windows servers? Disk space is cheap, so you may want to collect a large amount of data, but most of the time you really only need to look at a handful of metrics – for memory, I/O and CPU – to keep an eye on the big picture for just about any Windows, Linux or Unix server. What’s most important is that you look at them in that order – MIC – first Memory, then I/O, then CPU - because a performance issue in memory can contribute to I/O and CPU bottlenecks, and so on.

Remember: Memory First

For Windows servers, look at the system’s page rate and free memory. The page rate measures the number of pages read into (and out of) memory from disk per second. Occasional spikes are normal, but over continuous periods of time (15 minutes or so) the value should be well under 150. Similarly, look at the amount of free memory on the system; if this is consistently under 15%, the system is most likely paging (which it shouldn't) and is experiencing a performance hit.

Next In: I/O

Here, you really need to look at both the disk queue and the disk busy percentage. Monitor each logical disk in your server. If the Disk Queue Length exceeds 2 for continuous periods (15 minutes to an hour or so) for each disk drive in an array, then you probably have an I/O bottleneck for that array.

The Disk Busy % should run less than 50%. If this counter exceeds 50% for continuous periods, your system may be experiencing an I/O bottleneck.

If you do have an I/O bottleneck, look at which processes are consuming large amounts of CPU time; chances are, those processes are contributing to your I/O load.

Last but Not Least: CPU

If the CPU Busy % exceeds 80% for continuous periods (15 minutes or more), then you may have a CPU bottleneck on your server. Occasional spikes of 100% are nothing to worry about and are normal.

If you do suspect a bottleneck, take a look for processes using large amounts of CPU time. If you have one or a few CPU-intensive processes on the system, you will want to investigate possible bugs or efficiency problems with the application; or, it’s possible that high CPU utilization is simply a signature of this particular application.

High overall CPU usage without apparent CPU hogs may mean that you simply have too many concurrent users and may be reaching capacity.

As Always: Your Mileage May Vary

These are rules of thumb, and you will need to make adjustments based on the characteristics of your various expected workloads; however, if you monitor these metrics consistently, you will develop a clear picture of exactly what mileage to expect given your particular situation, and you will know when it’s time to delve deeper and think about a tune-up.

Next month: Linux Server Monitoring for Windows Admins