17 min of video discussing the balance between performance and capacity.
In this video Bryan Semple from vKernel discuss about performance and capacity. The video is quite long – 17 minutes. What's important and what's not. The video has been done in Toronto VMUG and the interaction with the users is very interesting since they brings in some points from their environments which might fits and be interesting for you.
A quick quote from vKernel's blog where I found the info:
The picture here is much as we have seen at other VMUgs. The low hanging fruit for cost savings is starting to go away. Yet the cost pressure is still there. Throwing more hardware at the problem is not the right answer. First you need to tame the performance issue, then optimize to bring costs down. The futility of eyeball analytics was pretty evident in many people’s reactions.
In larger environments (from 100VMs up), you loose the “visibility” of what VM has the right amount of Memory or right number of vCPU and it can also vary if your production needs changes from time to time.You say why? It's because there is too many metrics to look after for too many VMs….
Also, even if the initial configuraiton of your VMs is done right, the needs can change. Being able not only to monitor, but also change the configurations automatically of your VMs by adding/removing Memory or CPU during periodical checks. That's what tools available today from vKernel are providing.