Node termination support?

[hairyhenderson]

I'm wondering if it'd be appropriate to extend chaoskube to support terminating nodes as well as pods?

I think outright terminating VMs is probably definitely outside the scope of what this project wants to do (it feels better to keep it cloud-agnostic), but it may be useful to randomly cordon & drain nodes. If the Cluster Autoscaler is also being used, it will take care of terminating the empty nodes.

Feel free to close this if this is totally outside chaoskube's scope!