Vultr Kubernetes Engine (VKE): Beginner Setup Guide 2026

This post contains affiliate links. If you purchase through our links, we may earn a commission at no extra cost to you.

Managed Kubernetes on AWS EKS costs $73/month before you spin up a single worker node. Vultr Kubernetes Engine gives you the same CNCF-certified control plane for free — you only pay for the compute you actually run. That’s a meaningful difference if you’re learning Kubernetes or running workloads that don’t need an enterprise-grade ecosystem.

vultr

This guide walks you through the full setup from zero: creating a VKE cluster on the dashboard, connecting kubectl, deploying a real NGINX app, and exposing it with a Vultr Load Balancer. If you’ve never touched Kubernetes before, that’s fine — the VKE dashboard makes the first steps genuinely approachable. By the end you’ll have a running cluster and understand the key trade-offs before you commit to a plan.

What Is Vultr Kubernetes Engine?

VKE is Vultr’s fully managed Kubernetes service. Vultr handles the control plane — the API server, etcd (encrypted and backed up), and cluster networking — and you manage your workloads. Worker nodes are standard Vultr Cloud Compute instances, so if you’ve already used Vultr for a VPS, you’ll recognize the same server plans and pricing structure.

The service is certified by the Cloud Native Computing Foundation (CNCF) Kubernetes Conformance Program, which means it passes the full upstream test suite and works with standard Kubernetes tooling. As of June 2026, VKE supports Kubernetes v1.36 (latest), v1.35, and v1.34. The default container network interface is Calico.

One thing VKE is not: a one-click hosting platform. You’re still responsible for deploying an ingress controller, managing TLS certificates, and upgrading worker nodes. If you’re not ready for that level of ownership yet, a plain Vultr VPS with Docker is a reasonable stepping stone. our Docker on Vultr guide covers that path.

VKE Pricing: What You’ll Actually Pay

The control plane is free. You pay only for worker nodes, which bill at standard Vultr Cloud Compute hourly rates. Worker nodes require a minimum of 2 GB RAM.

High Performance Plans (AMD/Intel NVMe — recommended for Kubernetes):

Prices last checked: June 21, 2026.

The cheapest viable single-node cluster for learning is $12/month (1 vCPU / 2 GB). For a production-grade 3-node cluster at 4 GB RAM per node, you’re looking at $72/month ($24 × 3). You’ll also pay for any Load Balancers and Block Storage you attach — more on that below.

New Vultr accounts can get $300 in free credit for 30 days, which gives you plenty of runway to test VKE without committing. vultr

For a full breakdown of Vultr’s plan tiers and where VKE fits into the overall cost picture, see our Vultr pricing breakdown.

How to Create a VKE Cluster (No CLI Required)

The dashboard-driven setup is legitimately beginner-friendly. Here’s the full flow.

Step 1: Log In and Navigate to Kubernetes

1. Log in to console.vultr.com
2. In the left sidebar, click Products → Kubernetes
3. Click Add Cluster

Step 2: Configure the Cluster

Fill in the creation form:

• Cluster Name — pick something descriptive (my-first-cluster is fine for testing)
• Kubernetes Version — select from the dropdown; v1.36 is the latest as of June 2026
• High Availability — optional; enables multiple control plane replicas. Skip it for a learning setup
• Location — pick the region closest to your users or yourself
• Vultr Firewall — optional; you can attach a firewall group to restrict inbound traffic

Step 3: Configure the Node Pool

Under Cluster Configuration:

• Instance Type — choose High Performance for the best price/performance on Kubernetes workloads
• Node Plan — the $12/month 1 vCPU / 2 GB plan is fine for learning; go to $24/month (2 vCPU / 4 GB) if you want to run more than a few pods reliably
• Node Count — start with 1 node for testing; use 3 for anything production (odd numbers matter for etcd quorum)
• VPC Network — optional; attach a legacy VPC if you need private networking (see the limitation note below)

Step 4: Deploy and Wait

Click Deploy Now. Provisioning takes 3–5 minutes. You’ll see the cluster status change from “Pending” to “Running.”

Step 5: Download Your kubeconfig

1. Click on the cluster name once it shows “Running”
2. Click Download Configuration — this saves a .yaml file (usually named vke.yaml)

Keep this file somewhere safe. It contains credentials for your cluster.

Connect kubectl to Your VKE Cluster

kubectl is the command-line tool for interacting with any Kubernetes cluster. You’ll need it installed locally.

Install kubectl

Linux (snap):

sudo snap install kubectl --classic

Linux (manual):

curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
chmod +x kubectl
sudo cp kubectl /usr/local/bin/

macOS (Homebrew):

brew install kubectl

Windows (winget):

winget install -e --id CNCF.kubectl

Verify the installation with kubectl version — you should see the client version printed without errors.

Point kubectl at Your VKE Cluster

mkdir -p ~/.kube
cp ~/Downloads/vke.yaml ~/.kube/config

Then verify the connection:

kubectl get nodes
kubectl cluster-info

A healthy cluster returns your node(s) with Ready status. If you get an authentication error, double-check the kubeconfig path and that the cluster is fully running in the Vultr dashboard.

Essential kubectl Commands to Know

# List all pods across all namespaces
kubectl get pods --all-namespaces

# View logs from a specific pod
kubectl logs <pod-name>

# Describe a resource for debugging
kubectl describe node <node-name>

# Apply a manifest file
kubectl apply -f deployment.yaml

# Scale a deployment
kubectl scale deployment <name> --replicas=3

# Remove resources
kubectl delete -f deployment.yaml

The VKE connection setup takes about 2 minutes once the cluster is running — the kubeconfig download and kubectl get nodes flow is about as smooth as it gets for a managed Kubernetes provider.

Deploy Your First App: NGINX with a Load Balancer

This is the part where it gets real. The following manifest creates an NGINX deployment with 2 replicas and exposes it via a Vultr Load Balancer.

Create a file called deployment.yaml with this content:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-demo
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:latest
        ports:
        - containerPort: 80
---
apiVersion: v1
kind: Service
metadata:
  name: nginx-lb
  annotations:
    service.beta.kubernetes.io/vultr-loadbalancer-protocol: "http"
spec:
  type: LoadBalancer
  selector:
    app: nginx
  ports:
  - port: 80
    targetPort: 80

Apply it:

kubectl apply -f deployment.yaml

Check the deployment status:

kubectl get pods
kubectl get svc nginx-lb

The nginx-lb service will initially show <pending> in the EXTERNAL-IP column. Give it 1–2 minutes — Vultr’s Cloud Controller Manager provisions a load balancer in the background and links it to your worker nodes. Once the IP appears, paste it into your browser and you’ll see the default NGINX welcome page.

How the load balancer integration works: When you create a Service with type: LoadBalancer, VKE’s Cloud Controller Manager (CCM) automatically provisions a Vultr Cloud Load Balancer and connects it to your cluster nodes. You pay for the load balancer at standard Vultr rates — budget around $10–12/month for a basic load balancer on top of your node costs.

What VKE Doesn’t Include: Ingress Controller

VKE does not ship with a pre-configured ingress controller. For basic testing with a single service, the LoadBalancer type works fine. But if you need path-based routing, host-based routing, or TLS termination across multiple services, you’ll need to install NGINX Ingress Controller or Traefik manually, then add cert-manager with Let’s Encrypt for automatic SSL. That’s a separate guide — for now, the LoadBalancer approach gets your first app running.

VKE Limitations Worth Knowing Before You Commit

No managed service is without trade-offs. Here’s what VKE doesn’t do that you might expect:

No VPC 2.0 support. This is the biggest networking constraint. VKE clusters can only attach to Vultr’s legacy VPC networks, not VPC 2.0. If you’ve already migrated other infrastructure to VPC 2.0, your cluster can’t join that private network. Vultr hasn’t given a timeline for fixing this.

Worker node upgrades are manual. Vultr manages control plane upgrades, but you’re responsible for upgrading worker nodes through the “Manage Upgrades” tab. It’s not difficult, but it’s not automatic either.

No bare-metal workers. Worker nodes are virtual Cloud Compute instances only. Vultr’s bare metal servers can’t be added to a VKE node pool.

No built-in ingress controller. K3s ships with Traefik out of the box; VKE doesn’t. If you need HTTP routing beyond a single service, plan for the manual installation step.

These aren’t dealbreakers for most use cases, but they’re worth knowing going in.

VKE vs DigitalOcean DOKS vs Hetzner K3s: Which One?

Prices last checked: June 21, 2026.

The honest read on this table: Hetzner K3s is dramatically cheaper — roughly 60–70% less for worker nodes. But you’re trading cost for operational ownership: no managed control plane, no auto-healing nodes, manual certificate management. If you’re already doing DevOps for a living, that’s fine. If you’re a developer who wants to focus on your application, it’s a meaningful burden.

Between VKE and DOKS, the price is essentially identical. DigitalOcean’s UI is more polished and its App Platform integration makes it a better pick if you want a managed-everything experience. VKE is the right call if you’re already invested in the Vultr ecosystem — same account, same billing, familiar server plans. If you’re looking at this from scratch, check our Vultr vs DigitalOcean comparison for a deeper head-to-head on the platforms.

Node Pool Options and Autoscaling

VKE supports multiple node pools per cluster. The plan types worth knowing:

• High Performance (AMD/Intel NVMe) — best price/performance for Kubernetes; this is what I’d use for almost any workload
• High Frequency (Intel Xeon) — slightly better I/O; useful for database-heavy pods
• Regular Cloud Compute — cheaper but subject to CPU throttling on shared infrastructure; not recommended for K8s workloads
• Dedicated CPU — for CPU-intensive workloads where shared CPU would cause performance variance

VKE also supports the Cluster Autoscaler: set a min and max node count per pool, and the autoscaler adds or removes nodes based on pending pod resource requests. If a worker node is deleted accidentally outside the dashboard, VKE automatically redeploys a replacement to maintain the pool configuration.

FAQ

How much does Vultr Kubernetes Engine cost?

The control plane is free. You pay only for worker nodes at standard Vultr Cloud Compute rates — minimum $12/month for a 1 vCPU / 2 GB High Performance node. A 3-node production cluster at 4 GB RAM per node costs $72/month. Add ~$10–12/month per Load Balancer if your services need external access.

Is VKE good for production workloads?

It can be, with caveats. The free, CNCF-certified control plane and node self-healing are solid foundations. The limitations to plan around: no VPC 2.0 support, manual worker node upgrades, and no built-in ingress controller. For internal tooling, staging environments, or cost-sensitive production workloads, VKE is a reasonable choice. For mission-critical production with strict networking requirements, evaluate whether the VPC 2.0 limitation affects your architecture.

Does VKE support Kubernetes autoscaling?

Yes. VKE supports the Kubernetes Cluster Autoscaler — configure min/max node counts per pool and it scales nodes automatically based on pod resource requests. Horizontal Pod Autoscaler (HPA) for pod-level scaling also works with VKE like any standard Kubernetes cluster.

Can I use Helm with VKE?

Yes. VKE is CNCF-certified and fully compatible with Helm, Kustomize, and any standard Kubernetes tooling. Once your kubeconfig is in place, helm install works the same as on any other cluster.

What Kubernetes version does VKE support?

As of June 2026, VKE supports v1.36 (latest), v1.35, and v1.34. Vultr has been releasing updates regularly — the changelog shows multiple releases in 2026, including v1.36.1 in May 2026.

Final Verdict: Is VKE Worth It?

For developers who want managed Kubernetes without paying $73/month to AWS before running a single workload, VKE is a genuinely compelling option. The free control plane, CNCF certification, and node self-healing make it a real managed Kubernetes service — not a stripped-down toy. The dashboard setup is approachable enough that you can go from zero to a running cluster in under 10 minutes.

The VPC 2.0 limitation is real and worth thinking through if you have existing Vultr infrastructure. The lack of a built-in ingress controller means more setup work compared to K3s. And if polished UI and ecosystem integrations matter more than cost, DigitalOcean DOKS is worth a look.

But for a developer starting fresh on Kubernetes — especially one already using Vultr for other workloads — VKE at $12/month for a learning cluster is hard to argue with. New accounts get $300 in free credit, which is more than enough to run a full cluster for a month without spending anything.

vultr

Not ready for Kubernetes yet? A single Vultr VPS with Docker Compose covers most deployment needs at a lower operational overhead. Our Docker on Vultr guide walks through that setup. And if you’re still evaluating whether Vultr is the right platform overall, the Vultr Review 2026 covers the full platform in detail.