Skip to main content
This tutorial walks you through how to deploy a production-ready vLLM inference service on CoreWeave Kubernetes Service (CKS). By the end, you have a scalable LLM inference endpoint that runs the Llama 3.1 8B Instruct model on GPU Nodes. The endpoint is fronted by TLS-secured ingress and instrumented with Prometheus and Grafana for monitoring, plus KEDA for autoscaling. This tutorial is intended for platform engineers, ML engineers, and infrastructure operators who are comfortable with Kubernetes and want to serve large language models on CoreWeave GPU infrastructure.

​
Tutorial outline

This long-form tutorial comprises the pages underneath this section. Follow them in the order they are numbered, because each page builds on the resources deployed in the previous one. In this tutorial, you:
  1. Set up infrastructure dependencies.
  2. Configure monitoring and observability.
  3. Deploy vLLM inference service.
  4. Monitor performance and test autoscaling.

What you'll need

Before you start, you must have:
  • A working CKS cluster with GPU and CPU Nodes.
  • The following tools installed on your local machine:
    • Kubectl, connected to your cluster.
    • Helm version 3.8 or later.
    • Git.
  • Access to Hugging Face models (with tokens if required for gated models).

What you'll use

In this tutorial, you’ll use the following tools and services:
  • vLLM: High-performance LLM inference engine.
  • Traefik: Ingress controller for external traffic routing.
  • cert-manager: Automatic TLS certificate management.
  • Prometheus and Grafana: Monitoring and observability stack.
  • KEDA: Kubernetes Event-driven Autoscaling.
  • CoreWeave Helm charts: Preconfigured deployment templates.

​
Architecture overview

Before you start the procedures, it helps to understand how the pieces fit together. The complete vLLM inference solution consists of several components working together:
  • vLLM service: The main inference engine that runs your language model.
  • Traefik ingress: Handles external traffic routing and TLS termination.
  • cert-manager: Manages automatic SSL certificate generation and renewal.
  • Prometheus: Collects metrics from vLLM and other components.
  • Grafana: Provides dashboards that monitor inference performance.
  • KEDA: Enables autoscaling based on custom metrics like request queue depth.
Architecture diagram showing vLLM service on GPU Nodes with Traefik ingress, cert-manager, Prometheus, Grafana, and KEDA autoscaling.

​
Prerequisites

Before you begin the tutorial, verify that your environment meets the following requirements. Each check confirms a capability that later steps depend on.
  • You can access your cluster using kubectl. For example, run the following command: 
    kubectl cluster-info
    You should see something similar to the following: 
    Kubernetes control plane is running at...
CoreDNS is running at...
node-local-dns is running at...
  • Your cluster has at least one CPU Node. For example, run the following command: 
    kubectl get nodes -o=custom-columns="NAME:metadata.name,CLASS:metadata.labels['node\.coreweave\.cloud\/class']"
    You should see something similar to the following: 
    NAME      CLASS
g137a10   gpu
g5424e0   cpu
g77575e   cpu
gd926d4   gpu
  • Your CKS cluster must have GPU Nodes with at least 16 GB of GPU memory, which is required by the Llama 3.1 8B Instruct model used in this tutorial. For example, run the following command: 
    kubectl get nodes -o=custom-columns="NAME:metadata.name,IP:status.addresses[?(@.type=='InternalIP')].address,TYPE:metadata.labels['node\.coreweave\.cloud\/type'],RESERVED:metadata.labels['node\.coreweave\.cloud\/reserved'],NODEPOOL:metadata.labels['compute\.coreweave\.com\/node-pool'],READY:status.conditions[?(@.type=='Ready')].status,GPU:metadata.labels['gpu\.nvidia\.com/model'],VRAM:metadata.labels['gpu\.nvidia\.com/vram']"
    You should see something similar to the following: 
    NAME      IP               TYPE         RESERVED   NODEPOOL    READY   GPU           VRAM
g80eac0   10.176.212.195   gd-1xgh200   cw9a2f     infer-gpu   True    GH200_480GB   97
gf2809a   10.176.244.33    turin-gp-l   cw9a2f     infer-cpu   True
    Under VRAM, the number should be 16 or greater.
    To further debug and diagnose cluster problems, use kubectl cluster-info dump.
After these checks pass, you’re ready to begin the first page of the tutorial.

​
Additional resources and information

You don’t need to install the following tools yourself. They are preinstalled on CKS worker Nodes:
  • Docker: Container runtime for running vLLM inference Pods.
  • NVIDIA drivers: GPU drivers for CUDA acceleration.
  • CoreWeave CSI drivers: Storage drivers for persistent volumes.
  • CoreWeave CNI: Network plugins for Pod communication.
To learn more about vLLM and inference on CoreWeave, see the following resources:
Last modified on June 10, 2026