For those using PCF, how are you doing your allocation - are you burdening the Server object or the Application directly?  Were initially thinking of just sending the cost straight to the apps using it, but now not sure...  

In our case the PCF costs are appearing via the Cloud Cost Management module from Azure.  We bring the costs to Applications.  Our theory (for what its worth) is that this typically needs closer attention to manage costs and making the costs very visible to application owners is important.  With our servers object, this is basically on-prem costs that vary less.

Thank you, Sandor!

Update.. For right now - since we're in R11 and do not have CCM running, we're going to do a quick/dirty cost driver from CS to ITRT and from ITRT to the supported apps.  This will send the PCF related costs to the app.  We have various ESX clusters where we know the hypervisor and VM relationship and will use that to derive 'indirect' costs per se in a report later on.  So direct would be the invoices hitting the app and for indirect, we would layer on any database, storage, network associated w/the fully-burdened VMs/ESXs.  This is clearly a band-aid approach until we get going with R12/CCM.