Hyperstack On-demand Kubernetes: How to Scale Your Cluster Manually

Are you building cutting-edge AI applications or running distributed systems? Let Hyperstack streamline your infrastructure setup with on-demand Kubernetes so you can focus on innovation. Hyperstack’s On-Demand Kubernetes is a powerful, API-driven solution for provisioning Kubernetes clusters for modern AI workloads. With just a single API call, you can spin up a complete Kubernetes cluster to deploy and manage containerised applications. You can also scale it as your workload grows with our new APIs for cluster scaling. Scaling a Kubernetes cluster ensures your applications remain responsive and performant as workloads increase. 

Want to know how it works? Check out our quick tutorial below to understand how to scale Hyperstack On-demand clusters manually.

Why Choose Hyperstack On-demand Kubernetes?

Our On-Demand Kubernetes clusters are equipped with the following features, ensuring that your clusters are fully prepared to handle data-heavy AI workloads right from the start.

• Single API Request: You can deploy a complete Kubernetes cluster with just one API call, instantly provisioning essential components like the master node, load balancer, bastion VM and worker nodes.
• AI-Optimised Kubernetes: Built for Hyperstack’s infrastructure with custom NVIDIA-optimised drivers to enhance performance for your AI applications.
• Streamlined Workflows: Hyperstack’s intuitive APIs simplify deployment and automation, so you can manage and scale your Kubernetes clusters effectively.
• Effortless Deployment: You can launch fully configured clusters with minimal effort, as Hyperstack’s backend automates provisioning, enabling your clusters to run in minutes.
• High-Speed Networking: Hyperstack Kubernetes clusters use low-latency, high-speed networking of up to 350Gbps for distributed AI applications requiring fast data throughput.
• Role-Based Access Control (RBAC): You can secure and manage access with RBAC, ensuring your team has the appropriate permissions for seamless collaboration on AI projects.

Check out how to Set Up Kubernetes Clusters on Hyperstack: Step-by-Step Guide.

How to Scale Your On-Demand Kubernetes Cluster (Manually)

We’ve introduced scaling for Hyperstack’s On-Demand Kubernetes clusters, here’s how you can scale them manually to meet your growing demands in just two easy steps. 

Retrieve Cluster Node Details

GET: https://infrahub-api.nexgencloud.com/v1/core/clusters/{id}/nodes 

1. Retrieve the cluster ID by calling the List Clusters API.

2. View the node details of your Kubernetes cluster by including the cluster ID in the request path.

3. Refer to the full API documentation, including descriptions of the response fields, by clicking here.

Example Request

curl -X GET "https://infrahub-api.nexgencloud.com/v1/core/clusters/{id}/nodes" \
 -H "accept: application/json"\
 -H "api_key: YOUR API KEY"

Adding a Cluster Node

POST: https://infrahub-api.nexgencloud.com/v1/core/clusters/{id}/nodes 

1. Expand your existing Kubernetes cluster by adding one or more worker nodes. 
2. Include the cluster-ID in the request path and specify the desired node count in the request body. 
3. Ensure the role field is set to "worker," as only worker nodes can be dynamically added. For complete API details, click here.

Example Request

curl -X POST "https://infrahub-api.nexgencloud.com/v1/core/clusters/{id}/nodes" \
  -H "accept: application/json" \
  -H "api_key: YOUR_API_KEY" \
  -H "content-type: application/json" \
  -d '{
    "role": "worker",
    "count": { # of nodes}
  }'

Deleting a Cluster Node

DELETE: https://infrahub-api.nexgencloud.com/v1/core/clusters/{id}/nodes/{node_id} 

1. Obtain the node ID by calling the Retrieve Node Details API.
2. Remove a node from your Kubernetes cluster by including both the cluster ID and the node ID in the request path.
3. Access the full API documentation for additional details by clicking here.

Before Deleting a Node

Before deleting a node from your Kubernetes cluster, you must first drain and remove it from Kubernetes to prevent orphaned resources. Use the following kubectl commands to safely detach the node from Kubernetes:

kubectl drain <node-name> --ignore-daemonsets --delete-emptydir-data  # Remove running workloads from the node
kubectl delete node <node-name>  # Remove node from Kubernetes API

Once the node has been removed from Kubernetes, you can proceed with deleting it in Hyperstack via the API below.

Example Request

curl -X DELETE "https://infrahub-api.nexgencloud.com/v1/core/clusters/{id}/nodes/{node_id}" \
  -H "accept: application/json"\
  -H "api_key: YOUR API KEY" 

Important: Any data stored on a deleted node will be permanently lost and cannot be recovered. 

Conclusion

Hyperstack’s On-Demand Kubernetes simplifies scaling for modern AI workloads, offering a flexible, API-driven solution to manage clusters effortlessly. With features like single API call, NVIDIA GPU optimisation and high-speed networking, it’s built to handle data-intensive tasks while keeping your applications responsive. Manually scaling your cluster whether adding or removing worker nodes is an easy process with Hyperstack’s intuitive APIs, giving you precise control over resources. 

New to Hyperstack? Log in to Get Started with Our Cloud GPU Platform.

Explore Related Blogs

• Must-Know Tips for Managing Kubernetes Clusters

• Why Kubernetes is Essential for AI Workloads

• The Ultimate Guide to Kubernetes Architecture

• Kubernetes vs Docker: What is the Difference?

• How to Get Started with LLMs and Kubernetes

FAQs

What is Hyperstack On-Demand Kubernetes?

It’s an API-driven service to deploy and scale Kubernetes clusters for AI and data-intensive workloads. Launch clusters effortlessly with a single call.

How do I scale my cluster manually?

You can use Hyperstack’s scaling APIs to add or remove worker nodes by specifying the cluster ID and node count.

Can I add nodes other than worker nodes?

Currently, only worker nodes can be added dynamically; other roles are pre-configured during deployment.

What happens to data when I delete a node?

All data on a deleted node is permanently lost, so back up critical information first.

How fast can I deploy a Hyperstack On-demand Kybernetes cluster?

Clusters are fully operational within minutes, thanks to Hyperstack’s automated provisioning.

Is Hyperstack optimised for AI workloads?

Yes, it includes NVIDIA-optimised drivers and high-speed networking tailored for AI/ML tasks.