We’re implementing a Remote deployment of Cloud. The server requirements are very large. How would I go about deploying the Kubernetes cluster across multiple servers, not for fail-over purposes or for user-based load balancing but rather for load balancing the different pods?
Page 1 / 1
This is a really interesting topic. In Kubernetes there’s something called Horizontal Pod Autoscaling (HPA), which I think is what you might be talking about.
https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale/
The question is a bit unclear…
The HW requirements are high, so you want many smaller servers instead?
Or do you want more replicas of each pod?
or do you as maheshmuz suggest want to scale pods dynamically. Then also the cluster need to extend dynamically with node autoscaling. (For Remote installations that is not possible - there we have a fix number of nodes)
The HW requirements are high, so you want many smaller servers instead?
That’s what I’m exploring, yes. Can it be done?
I don’t see why not, but there is an overhead on each node, the most efficient way to minimize the total amount of cpu/memory is to only have one big server.
Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.