Upbound Docs - AWS Deployment Guide

AWS Deployment Guide

Get started with Upbound Spaces. This guide deploys a self-hosted Upbound cluster in AWS.

Disconnected Spaces allows you to host managed control planes in your preferred environment.

Prerequisites

To get started deploying your own Disconnected Space, you need:

An Upbound organization account string, provided by your Upbound account representative
A token.json license, provided by your Upbound account representative
An AWS account and the AWS CLI

Important

Disconnected Spaces are a business critical feature of Upbound and requires a license token to successfully complete the installation. Contact Upbound if you want to try out Upbound with Disconnected Spaces.

Provision the hosting environment

Create a cluster

Configure the name and target region you want the EKS cluster deployed to.

export SPACES_CLUSTER_NAME=upbound-space-quickstart
export SPACES_REGION=us-east-1

Provision a 3-node cluster using eksctl.

cat <<EOF | eksctl create cluster -f -
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
metadata:
  name: ${SPACES_CLUSTER_NAME}
  region: ${SPACES_REGION}
  version: "1.29"
managedNodeGroups:
  - name: ng-1
    instanceType: m5.xlarge
    desiredCapacity: 3
    volumeSize: 100
    iam:
      withAddonPolicies:
        ebs: true
iam:
  withOIDC: true
  serviceAccounts:
    - metadata:
        name: aws-load-balancer-controller
        namespace: kube-system
      wellKnownPolicies:
        awsLoadBalancerController: true
    - metadata:
        name: efs-csi-controller-sa
        namespace: kube-system
      wellKnownPolicies:
        efsCSIController: true
addons:
  - name: vpc-cni
    attachPolicyARNs:
      - arn:aws:iam::aws:policy/AmazonEKS_CNI_Policy
  - name: aws-ebs-csi-driver
    wellKnownPolicies:
      ebsCSIController: true
EOF

Configure the pre-install

Set your Upbound organization account details

Set your Upbound organization account string as an environment variable for use in future steps

export UPBOUND_ACCOUNT=

Set up pre-install configurations

Export the path of the license token JSON file provided by your Upbound account representative.

# Change the path to where you saved the token.
export SPACES_TOKEN_PATH=""

Set the version of Spaces software you want to install.

export SPACES_VERSION=1.9.5

Set the router host and cluster type. The SPACES_ROUTER_HOST is the domain name that’s used to access the control plane instances. It’s used by the ingress controller to route requests.

export SPACES_ROUTER_HOST=""

Important

Make sure to replace the placeholder text in SPACES_ROUTER_HOST and provide a real domain that you own.

Install the Spaces software

Install cert-manager

Install cert-manager.

kubectl apply -f https://github.com/cert-manager/cert-manager/releases/download/v1.11.0/cert-manager.yaml
kubectl wait deployment -n cert-manager cert-manager-webhook --for condition=Available=True --timeout=360s

Install ALB Load Balancer

helm install aws-load-balancer-controller aws-load-balancer-controller --namespace kube-system \
  --repo https://aws.github.io/eks-charts \
  --set clusterName=${SPACES_CLUSTER_NAME} \
  --set serviceAccount.create=false \
  --set serviceAccount.name=aws-load-balancer-controller \
  --wait

Install ingress-nginx

Install ingress-nginx.

helm upgrade --install ingress-nginx ingress-nginx \
  --create-namespace --namespace ingress-nginx \
  --repo https://kubernetes.github.io/ingress-nginx \
  --version 4.7.1 \
  --set 'controller.service.type=LoadBalancer' \
  --set 'controller.service.annotations.service\.beta\.kubernetes\.io/aws-load-balancer-type=external' \
  --set 'controller.service.annotations.service\.beta\.kubernetes\.io/aws-load-balancer-scheme=internet-facing' \
  --set 'controller.service.annotations.service\.beta\.kubernetes\.io/aws-load-balancer-nlb-target-type=ip' \
  --set 'controller.service.annotations.service\.beta\.kubernetes\.io/aws-load-balancer-healthcheck-protocol=http' \
  --set 'controller.service.annotations.service\.beta\.kubernetes\.io/aws-load-balancer-healthcheck-path=/healthz' \
  --set 'controller.service.annotations.service\.beta\.kubernetes\.io/aws-load-balancer-healthcheck-port=10254' \
  --wait

Install Upbound Spaces software

Create an image pull secret so that the cluster can pull Upbound Spaces images.

kubectl create ns upbound-system
kubectl -n upbound-system create secret docker-registry upbound-pull-secret \
  --docker-server=https://xpkg.upbound.io \
  --docker-username="$(jq -r .accessId $SPACES_TOKEN_PATH)" \
  --docker-password="$(jq -r .token $SPACES_TOKEN_PATH)"

jq -r .token $SPACES_TOKEN_PATH | helm registry login xpkg.upbound.io -u $(jq -r .accessId $SPACES_TOKEN_PATH) --password-stdin

Install the Spaces software.

helm -n upbound-system upgrade --install spaces \
  oci://xpkg.upbound.io/spaces-artifacts/spaces \
  --version "${SPACES_VERSION}" \
  --set "ingress.host=${SPACES_ROUTER_HOST}" \
  --set "account=${UPBOUND_ACCOUNT}" \
  --set "authentication.hubIdentities=true" \
  --set "authorization.hubRBAC=true" \
  --wait

Create a DNS record

Important

If you chose to create a public ingress, you also need to create a DNS record for the load balancer of the public facing ingress. Do this before you create your first control plane.

Create a DNS record for the load balancer of the public facing ingress. To get the address for the Ingress, run the following:

kubectl get ingress \
  -n upbound-system mxe-router-ingress \
  -o jsonpath='{.status.loadBalancer.ingress[0].hostname}'

If the preceding command doesn’t return a load balancer address then your provider may not have allocated it yet. Once it’s available, add a DNS record for the ROUTER_HOST to point to the given load balancer address. If it’s an IPv4 address, add an A record. If it’s a domain name, add a CNAME record.

You are ready to create your first managed control plane in your Space.

Create your first managed control plane

With your kubeconfig pointed at the Kubernetes cluster where you installed the Upbound Space software, create a managed control plane:

cat <<EOF | kubectl apply -f -
apiVersion: spaces.upbound.io/v1beta1
kind: ControlPlane
metadata:
  name: ctp1
  namespace: default
spec:
  writeConnectionSecretToRef:
    name: kubeconfig-ctp1
    namespace: default
EOF

The first managed control plane you create in a Space takes around 5 minutes to get into a condition=READY state. Wait until it’s ready using the following command:

kubectl wait controlplane ctp1 --for condition=Ready=True --timeout=360s

Connect to your managed control plane

Connect to your managed control plane with the up ctx command. With your kubeconfig still pointed at the Kubernetes cluster where you installed the Upbound Space, run the following:

up ctx default/ctp1

This command updates your current kubecontext. You’re now connected to your managed control plane directly. Confirm this is the case by trying to list the CRDs in your managed control plane:

kubectl get crds

To disconnect from your managed control plane and switch back to your previous context:

up ctx -

Tip

Learn how to use the up CLI to navigate around Upbound by reading the up ctx command reference.