How does CPU scheduling work?

July 2, 2012 4.3k views
4 Answers
Each Droplet has a specific number of virtual cores associated with it. Larger Droplet sizes receive more virtual cores. When a physical node is underutilized all Droplets are allowed to spike their CPU utilization to take advantage of available CPU cycles.
How about if all Droplets on a physical node are using a lot of CPU? Is there a minimum guaranteed amount of CPU that my droplet will receive? e.g. assume 4 droplets on a quad-core 1 CPU machine. Will each droplet get at least 1/4, that is, 1 full physical core at a minimum?
God question, but not answered....
The CPU cores are a logical designation which represents how many logical cores you have and determines that amount of parallel processing that your droplet can initiate.

The CPU scheduling is handled by the KVM virtualization engine on the backend of the hypervisor and because it is segmented logically, not physically, it will dynamically allocate more CPU cycles to your droplet when they are available.
Have another answer? Share your knowledge.