Leaseweb’s Kubernetes Service simplifies the creation and management of enterprise-grade Kubernetes clusters, making it easier to deploy containerized applications. Leaseweb handles your cluster resources and automates routine Kubernetes administration and scaling tasks, thereby significantly reducing the operational burden and streamlining Kubernetes operations.
We’re excited to announce the availability of cluster node autoscaling for our Managed Kubernetes service. You can now use the Cluster Autoscaler to dynamically scale nodes based on workload demand making cluster administrators overcome the need to manually scale the resources node resources.
Automate your Clusters
Leaseweb Managed Kubernetes has now added one of Kubernetes’ most powerful features: Horizontal (cluster-level) autoscaling. Cluster Autoscaler automatically adjusts the size of a Kubernetes cluster’s node pools based on workload demand. When demand increases, the number of nodes is scaled up to meet those demands. When demand decreases, instead of leaving an excess of resources assigned to your cluster, the number of nodes is scaled down.
What is Autoscaling in Kubernetes?
Autoscaling is the ability of a system to automatically adjust its capacity to match demand. In Kubernetes, autoscaling can be applied both to the cluster level (the number of nodes) and to the pod level (the number of pod replicas). Autoscaling ensures that resources are utilized optimally, minimizing waste while maintaining application performance.
Pod-level scaling allows you to spin up multiple instances of an application, usually to either guarantee redundancy, or to make sure you have enough replicas to meet the load / demand.
However, pod-level scaling benefits are usually restricted by the total number of resources available in your cluster (number of nodes, total CPU, and memory), which must be tuned manually.
Cluster-level autoscaling allows your cluster to use more nodes during peak demands, rapidly granting more resources (CPU, Memory) to your pods, to allow you to meet your workloads’ demands.
How to configure the scaling of your clusters
You can configure the autoscaling by setting a minimum and maximum number of nodes. This allows you to control the range within which your cluster can scale, ensuring flexibility while preventing over-scaling.
Scaling Up (Adding Nodes)
It is recommended to have an autoscaling configured to adjust the number of pod replicas based on workload demands. When your application experiences increased demands, Kubernetes will automatically provision pods to meet your request. Once no more space is available, the newly spawned pods will be in a “pending” state, waiting for new resources to be added. Cluster Autoscaler will automatically scale up the cluster by adding more nodes. This ensures your applications have enough capacity to handle the surge in traffic or resource usage, without any manual intervention or overprovisioning.
Scaling Down (Removing Nodes)
As workloads decrease, Kubernetes periodically re-evaluates if your cluster can be reduced according to the configured range, and will automatically remove underutilized nodes from the cluster, ensuring you don’t pay for unused resources. The scaling process happens without disrupting running applications, as nodes are gracefully terminated once they are no longer needed.
Understanding the Cost implications
While configuring your cluster, we have also added support via the customer portal to project and help you plan cost and preview how the scale impacts price.
Contact us to learn how our combined kubernetes solution can transform your application infrastructure while reducing complexity and costs.