Module 1: Setup

ArgoCD on EKS demo

This demo covers Application deployment (both runtime and infrastructure services) and Addons management in a multi-cluster scenario, where a single Argo CD (hub) cluster manages the deployment to all other workload clusters (spokes) in the organization

Use Cases

Deploy EKS clusters (hub, staging, prod)
Deploy Namespaces
Create DynamoDB
Deploy Applications
Day 2 OPS

Module 1: Setup

Run the following command create git repository in CodeCommit and create 3 EKS Clusters (Hub, Staging, Prod)

export ACCOUNT_ID=$(aws sts get-caller-identity --output text --query Account)
export AWS_DEFAULT_REGION="eu-west-1"
export WORKING_DIR="$HOME/environment" #For local dev; export WORKING_DIR="$PWD/environment"
export WORKSHOP_DIR="$WORKING_DIR/argocd-eks" #For local dev; export WORKSHOP_DIR="$PWD"
export GITOPS_DIR="$WORKING_DIR/gitops-repos"
git clone https://github.com/moabukars/argocd-eks $WORKSHOP_DIR
cd $WORKSHOP_DIR
./install.sh

Access ArgoCD UI on the Hub Cluster

echo "ArgoCD URL: https://$(kubectl --context hub-cluster get svc -n argocd argo-cd-argocd-server -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')"
echo "ArgoCD Username: admin"
echo "ArgoCD Password: $(kubectl --context hub-cluster get secrets argocd-initial-admin-secret -n argocd --template="{{index .data.password | base64decode}}")"

Module 2: Platform

Deploy EKS Addons

To deploy EKS Addons we need to use Infrastructure as Code (IaC) and GitOps to work together. Enable the variables for the corresponding Addons in the IaC tool.

sed -i '' s/"balancer_controller = false"/"balancer_controller = true"/           $WORKSHOP_DIR/terraform/spokes/variables.tf
sed -i '' s/"dynamodb                 = false"/"dynamodb                 = true"/ $WORKSHOP_DIR/terraform/spokes/variables.tf

Apply the IaC to create the IAM Roles for each Addon, and enable the Helm Chart to be deploy by GitOps

terraform -chdir=$WORKSHOP_DIR/terraform/spokes workspace select staging
terraform -chdir=$WORKSHOP_DIR/terraform/spokes apply -var-file="workspaces/staging.tfvars" -auto-approve
terraform -chdir=$WORKSHOP_DIR/terraform/spokes workspace select prod
terraform -chdir=$WORKSHOP_DIR/terraform/spokes apply -var-file="workspaces/prod.tfvars" -auto-approve

Verify AWS Load Balancer Controller is installed on each cluster

kubectl get pods -n kube-system -l app.kubernetes.io/name=aws-load-balancer-controller --context staging-cluster
kubectl get pods -n kube-system -l app.kubernetes.io/name=aws-load-balancer-controller --context prod-cluster

Expected Result

NAME                                            READY   STATUS    RESTARTS   AGE
aws-load-balancer-controller-6d64b58dfc-89wxc   1/1     Running   0          3m
aws-load-balancer-controller-6d64b58dfc-pzbtl   1/1     Running   0          3m

Verify AWS Controller for Kubernetes (ACK) is installed on each cluster

kubectl get pods -l app.kubernetes.io/instance=ack-dynamodb -A --context staging-cluster
kubectl get pods -l app.kubernetes.io/instance=ack-dynamodb -A --context prod-cluster

Expected Result

NAME                                               READY   STATUS    RESTARTS   AGE
ack-dynamodb-aws-controllers-k8s-89685f8c5-p9rwt   1/1     Running   0          3m

Deploy Tenants/Teams

For each team (ie. backend, frontend), the following will be created

ArgoCD Project
Namespaces, Network Policies, LimitRanges, ResourceQuotas

cp -r $WORKSHOP_DIR/gitops/platform/* $GITOPS_DIR/platform/
git -C $GITOPS_DIR/platform add .
git -C $GITOPS_DIR/platform commit -m "add platform"
git -C $GITOPS_DIR/platform push

Verify Namespaces

kubectl get ns -l app.kubernetes.io/created-by=eks-workshop \
  -o custom-columns='NAME:.metadata.name,TEAM:.metadata.labels.app\.kubernetes\.io/name,ENVIRONMENT:.metadata.labels.environment' \
  --sort-by '.metadata.labels.environment' --context staging-cluster
kubectl get ns -l app.kubernetes.io/created-by=eks-workshop \
  -o custom-columns='NAME:.metadata.name,TEAM:.metadata.labels.app\.kubernetes\.io/name,ENVIRONMENT:.metadata.labels.environment' \
  --sort-by '.metadata.labels.environment' --context prod-cluster

There should be new namespaces on each cluster

NAME       TEAM       ENVIRONMENT
carts      backend    staging
catalog    backend    staging
checkout   backend    staging
orders     backend    staging
rabbitmq   backend    staging
assets     frontend   staging
ui         frontend   staging

Module 3: Workloads in Stage

Deploy the workloads to staging cluster (default)

cp -r $WORKSHOP_DIR/gitops/apps/* $GITOPS_DIR/apps/
sed -i '' "s/ACCOUNT_ID/$ACCOUNT_ID/" $GITOPS_DIR/apps/backend/carts/staging/kustomization.yaml
sed -i '' "s/ACCOUNT_ID/$ACCOUNT_ID/" $GITOPS_DIR/apps/backend/carts/prod/kustomization.yaml
git -C $GITOPS_DIR/apps/ add .
git -C $GITOPS_DIR/apps/ commit -m "add workloads"
git -C $GITOPS_DIR/apps/ push

Files are committed for both staging and prod, but prod is inactive. In a later section we will deploy to prod

Verify the Amazon DynamoDB Table in staging

kubectl -n carts get tables.dynamodb.services.k8s.aws --context staging-cluster

The expected output should have the STATUS colum with ACTIVE and SYNCED column with True

NAME    CLASS      STATUS   SYNCED   AGE
items   STANDARD   ACTIVE   True     78s

Verify App is running on staging cluster

kubectl get pods -A -l app.kubernetes.io/created-by=eks-workshop --context staging-cluster

There should be pods running in namespaces

NAMESPACE   NAME                             READY   STATUS    RESTARTS        AGE
assets      assets-7556557b4d-sfndh          1/1     Running   0               3m56s
carts       carts-86c7db99db-5b557           1/1     Running   0               69s
carts       carts-dynamodb-cb4b6f564-wn265   1/1     Running   0               69s
catalog     catalog-6ccfd94978-fcrkl         1/1     Running   0               3m56s
catalog     catalog-mysql-0                  1/1     Running   0               3m56s
checkout    checkout-6845d66fb-b6m82         1/1     Running   0               4m
checkout    checkout-redis-fb67f7944-j2jvr   1/1     Running   0               4m1s
orders      orders-548b6658ff-qjsv7          1/1     Running   0               3m56s
orders      orders-mysql-76dd47c48f-jwt7j    1/1     Running   0               3m56s
ui          ui-59b974ffcc-cbmkw              1/1     Running   0               69s

Access Application UI

Access in

# TODO we need script to wait for url to be OK/200
# wait-lb http://$(kubectl --context prod-cluster get svc -n ui ui-nlb -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')
echo "Staging UI URL: http://$(kubectl --context staging-cluster get svc -n ui ui-nlb -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')"

Module 4: Day 2 Operations

Platform team promotes the app to prod cluster, by updating the ApplicationSet workloads.yaml

sed -i '' s/"values: \[staging\]"/"values: \[staging,prod\]"/ $GITOPS_DIR/platform/bootstrap/workloads/*appset.yaml
git -C $GITOPS_DIR/platform add .
git -C $GITOPS_DIR/platform commit -m "add workloads"
git -C $GITOPS_DIR/platform push

Verify the Amazon DynamoDB Table in prod

kubectl -n carts get tables.dynamodb.services.k8s.aws --context prod-cluster

The expected output should have the STATUS colum with ACTIVE and SYNCED column with True

NAME    CLASS      STATUS   SYNCED   AGE
items   STANDARD   ACTIVE   True     78s

Verify App is running on prod cluster

kubectl get pods -A -l app.kubernetes.io/created-by=eks-workshop --context prod-cluster

There should be pods running in namespaces

NAMESPACE   NAME                             READY   STATUS    RESTARTS        AGE
assets      assets-7556557b4d-sfndh          1/1     Running   0               3m56s
carts       carts-86c7db99db-5b557           1/1     Running   0               69s
carts       carts-dynamodb-cb4b6f564-wn265   1/1     Running   0               69s
catalog     catalog-6ccfd94978-fcrkl         1/1     Running   0               3m56s
catalog     catalog-mysql-0                  1/1     Running   0               3m56s
checkout    checkout-6845d66fb-b6m82         1/1     Running   0               4m
checkout    checkout-redis-fb67f7944-j2jvr   1/1     Running   0               4m1s
orders      orders-548b6658ff-qjsv7          1/1     Running   0               3m56s
orders      orders-mysql-76dd47c48f-jwt7j    1/1     Running   0               3m56s
ui          ui-59b974ffcc-cbmkw              1/1     Running   0               69s

Access Application UI in Prod

Access in

# TODO we need script to wait for url to be OK/200
# wait-lb http://$(kubectl --context prod-cluster get svc -n ui ui-nlb -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')
echo "Production UI URL: http://$(kubectl --context prod-cluster get svc -n ui ui-nlb -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')"

Update Workload in Production

Make a change in production, like setting the ui HPA minReplicas to 3. Update the patch $GITOPS_DIR//apps/ui/prod/hpa.yaml

Push the changes to git

sed -i '' s/"minReplicas: 2"/"minReplicas: 3"/ $GITOPS_DIR/apps/frontend/ui/prod/hpa.yaml
git -C $GITOPS_DIR/apps add .
git -C $GITOPS_DIR/apps commit -m "set minReplicas to 3 in prod"
git -C $GITOPS_DIR/apps push

Verify HPA MINPODS is now 3 in prod

kubectl get hpa -n ui --context prod-cluster

NAME   REFERENCE       TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
ui     Deployment/ui   2%/80%    3         10        3          29m

Verify there are at least 3 replias in prod cluster

kubectl get pods -n ui --context prod-cluster

Expected output

NAME                  READY   STATUS    RESTARTS   AGE
ui-59b974ffcc-fktvx   1/1     Running   0          5m
ui-59b974ffcc-hwkqh   1/1     Running   0          41s
ui-59b974ffcc-996z8   1/1     Running   0          3m23s

Cleanup

Destroy all AWS resources

./cleanup.sh

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
gitops		gitops
terraform		terraform
.gitignore		.gitignore
README.md		README.md
cleanup.sh		cleanup.sh
install.sh		install.sh
setup-git.sh		setup-git.sh
setup-kubectx.sh		setup-kubectx.sh
text.txt		text.txt
tfsec.yaml		tfsec.yaml

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

ArgoCD on EKS demo

Use Cases

Module 1: Setup

Module 2: Platform

Deploy EKS Addons

Deploy Tenants/Teams

Verify Namespaces

Module 3: Workloads in Stage

Access Application UI

Module 4: Day 2 Operations

Access Application UI in Prod

Update Workload in Production

Cleanup

About

Releases

Packages

Languages

moabukar/argocd-eks

Folders and files

Latest commit

History

Repository files navigation

ArgoCD on EKS demo

Use Cases

Module 1: Setup

Module 2: Platform

Deploy EKS Addons

Deploy Tenants/Teams

Verify Namespaces

Module 3: Workloads in Stage

Access Application UI

Module 4: Day 2 Operations

Access Application UI in Prod

Update Workload in Production

Cleanup

About

Resources

Stars

Watchers

Forks

Releases

Packages 0

Languages

Packages