Four Rules For Choosing Virtualisation Metrics

Monitoring performance on a single system is complicated enough, but what should you do when you’re tracking virtual machines? Here are four basic principles to apply before choosing.

Measurement picture from Shutterstock

These suggestions come from a presentation at TechEd North America 2013 by Raymound Chou from Infront Consulting and Alec King from Veeam Software. While they’re not rocket science, that doesn’t mean everyone will remember them. An example of obliviousness to the obvious? Chou pointed out that some people still use Task Manager on the host system for a basic measure of server performance, a fruitless task since it will only measure the performance of the base OS, not all the others running on the same platform.

Map out your entire system “You need to be able to trace the main relationships and dependencies,” King said.

Visualise the key elements A visual representation is much easier to track and monitor.

Make sure the chosen metrics play nicely with other tools. Performance data should be easy to import into the tracking tool of your choice.

Make sure you can develop a plan. Use metrics to feed forecasting and identify trends.

Visit Lifehacker’s World of Servers Newsroom for all the latest news from TechEd North America 2013. And don’t forget: TechEd is coming to Australia in September. Click here for more information.