It really all depends,
First of you need to define what you classify as "overload" There are 2 methods there is demand and allocation. Demand is automatically calculated based on a number of metrics and will usually give you a good idea what is actually being demanded by the guest VMs.
Second is allocation this is what you can define as in 6:1 vCPU to pCPU this is based on what your requirements are. Being an administrator you usually have a good idea on how the cluster is running and what it limits have.
You can both enabled and go with the most restrictive
These are configured though the policies applied to the cluster
There is lots of good doco around for it.
Yeah i understand what u mean, i need to change allocation for a 2:1 vCPU