When autoscaling, adding instances can take time to deploy on the cloud provider’s infrastructure, this means that the user will have to wait a while to reach the desired new scaling capacity for the Pods. Luckily, there have been major improvements to help users scale more efficiently.
We are pleased to announce significant improvements to Elastigroup’s Kubernetes Autoscaling. Now clusters will be able to scale Pods and underlying infrastructure faster than ever. When a Pod is scheduled to scale, Elastigroup will scale the appropriate amount of instances for all unscheduled Pods at the first scale up event while considering all of the constraints, persistent volume claims, resources, taints and tolerations.
To make scheduling more efficient and compatible with Kubernetes, the Elastigroup supports all of the Kubernetes constraint mechanisms for scheduling pods:
- Node Selector – Constrain Pods to nodes with particular labels.
- Node Affinity – Constrain which nodes your Pod is eligible to be scheduled on based on labels on the node.
- Pod Affinity and Pod Anti-Addinity – Schedules a Pod based on which other Pods are or are not running on a node.
Learn more about our Kubernetes Autoscaling in this post.