posted

5 Comments

This is part 2 of the mini Tech Tip series for vSphere Capacity Planning. In part 1 of this series, we covered the importance of using the right metrics and translating your operational knobs around buffers/over-commit etc into policies before you assess your capacity risk.

In this second Tech Tip,  we will cover examples out of box and custom visuals you could use in vCenter Operations Manager  to answer your key capacity planning questions:
a) Analyze VM growth across your environment
b) Understand how your infrastructure is getting used up to support the above growth
c) As you grow, do you have capacity risk now or in the future?

a) Analyze VM growth across your environment
After logging into vCenter Operations’ vSphere User Interface,  click on Planning and under Views, click on VM Capacity VIew. This view tells your how your VM growth has been for the selected cluster/datacenter/vCenter etc on the left navigation pane.It also projects out when you will run out and how much time you have.

Some IT admins may want to report growth across your environments such has production/application, Geo-region etc. Here is a custom dashboard you could create using the scoreboard widget and metric graph to plot count of total, powered on/off VMs, split by cluster/region etc

 

b) Understand how your infrastructure is getting used up to support the above growth
How are your clusters being used up over time? Remember to plot usable capacity ( which takes into account HA/buffers etc) vs total available capacity. Split across clusters/regions etc to get a view of usage growth.

 

Perhaps you want to report on ‘Current Utilization’-
i) which clusters are at 50%  utilization? Or report on total/usage across clusters
ii) which clusters are under-sized and have VMs experiencing performance issues/contention?

Here is a custom dashboard that shows total capacity, usage, which hosts in the cluster are undersized for their expected peak usage periods, which VMs on these hosts are undersized for their peak workloads?

Which clusters are under-utilized and can be consolidated? Under vSphere UI –> click Planning –>  click Underused Hosts and Clusters View

c) Ok, I understand my growth, and how my infra is used up, but do I have capacity risk based on my policies? Why? where?
– am I out of capacity now? Which resource?
– will I run out of capacity soon?
– is my cluster under-sized for expected peak usage periods?
– or can I see what is my current VM/Host ratio and has it exceeded my policy in production?

Ensure you have the right policies set to assess capacity risk in Configuration Settings: Production Policy ( uses allocation approach to assess risk) vs Test-Dev (which uses Demand based approach to assess risk).

This view below is a really important view – It tells you immediately whether you are our of capacity because you have reached your target threshold of allocating capacity to all the VMs on that cluster or whether the demand (CPU or Mem) across all the VMs has exceeded the desired thresholds….
(The dashboard is updated daily and the views are updated real time after applying the policies)
Under vSphere UI –> click Planning –> click Cluster Capacity Risk View

Hope this was useful, don’t forget to check out the part 3 of 4 of this Tech Tip series on right sizing the VMs in your virtual infrastructure!