Hello Replicated Team – we are curious if the Replicated tools also provide a management capability for Kuberentes, similar to AWS Managed EKS. The goal for us is to have our NODES to autoscale.
Hi Tushar!
Today in the kURL project we don’t have a way to support node autoscaling. If I think about the use case for most of our vendors’ end-customers who are deploying an application with the kURL embedded cluster, those folks usually already have “on-prem” or existing resources that they want to utilize, like perhaps a VMWare cluster or bare-metal machines. And we tend to find that those customers don’t necessarily have a trivial way to create new VMs or even new hard disks or firewall rules in those environments – they might have to go through a change management process to get those resources. So dynamic creation of VMs on customer-provided infrastructure is probably going to have lots of edge cases that may be difficult to predict or handle ahead of time.
Managed clusters like Azure, AWS and GCP already support node autoscaling based on app requirements, so if a customer wants a managed cluster I think these are fine solutions for that. Right now I have several applications deployed to a GCP Autopilot managed Kubernetes cluster, and it uses the Kubernetes requests.cpu and requests.memory keys to scale the size of the cluster appropriately. If a customer has a requirement that they want to support node autoscaling, we recommend they use one of the managed cluster providers that have that functionality.
If you want to talk in more depth about your design goals, though, send a ticket to our support team via the vendor portal at vendor.replicated.com and let’s have a chat.
Cheers,
Ada
Hello – yes will open a support ticket and thanks for the update