We experience this as well in our environment. The problem as I understand it is we really need a stub for the "Decommission" state of the lifecycle. In the realm of 5.x and lower I believe if you had the sdk you could pretty easily set this up. Unfortunately the way it stands today we have a limited set of those lifecycle callouts we can use. I think that you can in fact change the state at which the workflow stub gets called by modifying a file on the IaaS server (c:\Program files (x86)\VMware\vCAC\Server\ExternalWorkflows\xmldb\ExternalWFStubs.xml) . I'd imagine this is likely unsupported so if you try it be careful. I have not attempted this myself so again proceed with caution! You will need to modify the stub mapping for MachineDisposing to actually trigger at a the desired state.
I have asked about a million times now for all the states and sub-states to be made available as I can imagine making use of just about all of them for something. I have been told enhancements are coming. Doubtful we will see all of the available status types surface but surely a "pre-disposing" stub has got to be high on everyones list.
Here is a good article on the states. http://dailyhypervisor.com/vcloud-automation-center-vcac-5-2-virtual-machine-lifecycle-demystified/#... . It is written about 5.2 but as far as I can tell its all relevant to 6.x as well.
Paul
