Deploy AWX on k3s Single Node for Ansible Automation

Dec 29, 2025

Fadhli Ari Pratama

Open-Source Enthusiast

There comes a point in almost every infrastructure or DevOps engineer’s journey when Ansible CLI alone no longer feels enough.
Playbooks keep growing, inventories get more complex, credentials become sensitive, and eventually a question appears:

“I need a UI, RBAC, and job history — but I don’t want to run a full-blown Kubernetes stack.”

That’s where AWX starts to make sense.
And to keep things lightweight, k3s becomes a very reasonable foundation.

This article is not an enterprise best practice guide.
It’s a practical lab note — simple, fast, tidy enough, and realistic for daily use.

In many teams, the initial requirement is simple: “We need Ansible, but with a UI, RBAC, and job history.”

Officially, the answers are:

• Ansible Tower (now deprecated)
• Red Hat Ansible Automation Platform (AAP)

The challenge is:

• AAP is commercial
• Pricing can be overkill for labs or small internal teams
• Enterprise support is not always a strict requirement

This is where AWX becomes a pragmatic choice:

• Open-source
• Upstream project of Ansible Tower
• Core features like UI, RBAC, scheduling, and logging are available
• Can run on lightweight Kubernetes distributions such as k3s

If you require official support, strict compliance, and SLAs, Red Hat AAP remains the right choice.
But if flexibility and functionality are the priority, AWX is often more than enough.

The goal of this setup is intentionally modest:

• Single-node Kubernetes
• Fast installation
• Proof-of-concept environments
• No complex ingress configuration
• AWX accessible immediately
• Suitable for homelab or internal automation
• Small teams that need control without enterprise overhead

Not HA. Not enterprise-grade. But usable.

• Ubuntu Server 22.04
• 1 VM or bare metal node
• Minimum resources:
  • 2 vCPU
  • 4 GB RAM (8 GB recommended)
• Internet access

Start with the essentials. Nothing fancy, but explicit.

sudo apt update
sudo apt install -y git build-essential curl

One of the reasons k3s is appealing:

• Single binary
• Bundled container runtime
• Built-in Traefik
• Built-in local storage provisioner

Installation is literally one command:

curl -sfL https://get.k3s.io | INSTALL_K3S_EXEC="--write-kubeconfig-mode=644" sh -

The --write-kubeconfig-mode=644 flag allows kubectl usage without sudo.

Verify the node:

kubectl get nodes -o wide

Proceed once the node is Ready.

This step is often skipped, but it’s useful as a sanity check.

kubectl get all -n kube-system

You should see components such as:

• coredns
• metrics-server
• local-path-provisioner
• traefik
• svclb-traefik

Seeing Traefik assigned an external IP from your LAN subnet is expected behavior in k3s.

AWX is deployed exclusively via an operator.

Clone the official repository:

git clone https://github.com/ansible/awx-operator.git
cd awx-operator

Checkout a stable version:

git checkout tags/2.19.1
export VERSION=2.19.1
make deploy

The detached HEAD warning is expected when checking out a tag.

Verify the operator:

kubectl get all -n awx

Ensure the controller manager pod is in Running state.

Before deploying the AWX instance, the admin password must exist.

kubectl create secret generic awx-admin-password \
  --from-literal=password='<STRONG_PASSWORD>' \
  -n awx

Confirm:

kubectl get secrets -n awx

Create the manifest file awx-prod.yml:

apiVersion: awx.ansible.com/v1beta1
kind: AWX
metadata:
  name: awx-production
  namespace: awx
spec:
  admin_user: admin
  admin_password_secret: awx-admin-password
  service_type: nodeport

Apply the manifest:

kubectl apply -f awx-prod.yml
kubectl get pods -n awx -w

At this stage, several pods will be created, including PostgreSQL, awx-web, and awx-task.
This is the most resource-intensive part — give it some time.

Retrieve the NodePort service:

kubectl get svc -n awx

Access AWX using:

http://<NODE_IP>:<NODEPORT>

Login credentials:

• Username: admin
• Password: value from the secret

This setup is:

• ❌ Not highly available
• ❌ Not enterprise production-ready
• ✅ Fast to deploy
• ✅ Lightweight
• ✅ Ideal for labs and internal usage

It works well if you:

• Are transitioning from Ansible CLI
• Need RBAC and job history
• Want visibility without full Kubernetes complexity

For a starting point, however, this setup is more than sufficient.

AWX on k3s is an underrated combination.
It’s not for everyone, but it fits engineers who value pragmatism over complexity.

If you need automation that actually gets used, this setup is worth trying.

Happy automating ☕