Basically the Sunburst charts appear to misleading when over-subscription is taking place, both their colours and figures presented when looking at hosts and cluster segments.
In our test environment we had configured a Sunburst chart for a Dev Cluster which has the following properties.
Hosts # Cores Freq(GHz) Cores per socket CPU Size (MHz)
s2tesx1 12 3.4 6 40752
s2tesx11 12 2.5 6 29988
s2tesx12 12 2.5 6 29988
s2tesx13 12 2.5 6 29988
s2tesx16 12 2.5 6 29988
s2tesx2 12 3.4 6 40752
If I begin to look at how the Sunburst chart is displayed and the details presented for VM, ESX and Cluster, here's a VM ( last 30 days ) with it's segment correctly shown as red.
Segment colours for VMs appear to be generally correct.
The small "Details" pop-up shows utilisation of 152.15% yet the CPU Utilisation graph shows ~60%.
Why the difference between the two figures? Is the 152.15% a "demand" figure?
If I now look at the ESX host s2tesx16 ( 30GHz capacity ) the segment is blue in colour, shows utilisation of 14.58% and size of 247.4GHz, yet the graph to the side indicates ~ 80%.
1. The blue colour for the segment is wrong
2. The 14.58% utilisation in the pop-up is wrong - see the graph to the side
3. Does the size of 247.4GHz indicate over-subscription, in which case this figure and perhaps an over-subscription ratio figure ( e.g. 5:1 ) could be provided in the pop-up and possibly making the segment yet another colour ( black flashing !! ) - but definitely not blue.
Finally if I look at the Site 2 Dev Cluster circle it shows as blue with utilisation of 19.73% and size of 1.18 THz ( I wish !! ) whereas the CPU utilisation graph indicated just below 100%.
The same three points mentioned above would apply to the Cluster display information and colour.
So my request would be, in the even of Virtualisation over-subscription, to present Sunburst chart colours and figures more accurately.