In my previous post for this series I touched on both the allocation and demand models available to us when using vCops for capacity planning. As we have seen, playing with vCops capacity planning policy is important. Let’s continue with the configurations.
Tip #2 – The importance of usable capacity and over-commitment ratios
A big issue with vCops capacity module is getting realistic result. Like with creating a vSphere design, taking infrastructure buffers, over-commitment ratios and HA admission control policy under consideration for the environment policy, will get you closer towards real life results in terms of capacity planning. In some environments like Test/Dev you may not care about those stuff but in your production environment you should.
Example of high performance production environment usable capacity policy:
- In Planning cycle mode vCops looks at historical data based on business cycle defined by the number in “days of provisioning buffer”
- In Real time mode vCops looks at real time data and will not take historical data in its calculations
Example of highly utilize Dev environment usable capacity policy:
Now that you have a decent understanding on how to configure the ratios, buffers and time constrains (if there are any) you can go and match right policy for the right environment.
Under the ‘Usable Capacity’ how do you decide your buffer?
It is really up to you and your org policy.
If your org has a policy saying “once we hit the 90% datastore used space threshold no more VMs will be deployed” this is something to take under consideration and the outcome will be 10% buffer for disk space capacity. That goes for each one of rules available to you.
If you don’t have that kind of policy, just use your VI admin instincts 🙂