Search⌘ K
AI Features

Multiple Schedulers

Explore how Kubernetes handles multiple schedulers by assigning Pods to different schedulers using scheduler names. Learn how to configure multiple schedulers, the operational challenges they bring, and why maintaining custom schedulers requires deep knowledge of Kubernetes scheduling mechanics.

Multiple schedulers

In Kubernetes, a scheduler works similar to an operator or controller. It watches all the Pods and nodes, finds the nodes of best fit for newly created Pods, and binds node info to the Pod spec.nodeName. As a result, to meet our special scheduling requirements, one straightforward way to extend Kubernetes scheduling is by running a separate customized scheduler along with the default scheduler. We can run multiple such schedulers simultaneously.

How multiple schedulers work together

Each scheduler should have a different name to indicate itself. We instruct Kubernetes which scheduler to use for a Pod based on the spec.schedulerName in the Pod specification. Schedulers having the matching name with Pod spec.schedulerName are responsible for scheduling ...