Ingress Controller Change

This page outlines the steps to migrate from ingress-nginx to the F5 NGINX Ingress Controller.

Background

Following the Ingress NGINX Retirement in March 2026, kdb Insights Enterprise has been updated to ensure continued use of a supported ingress controller. Both the product and the kxi-terraform scripts now support the F5 NGINX Ingress Controller.

Prerequisites

The F5 NGINX Ingress / nginx-community controller is included in the kxi-terraform scripts for kdb Insights Enterprise version 1.18.3 and later.

To install the new controller in an existing deployment, complete the following steps:

Download the kxi-terraform bundle version 1.18.3 or later.
Untar into a new directory on the host used to run the terraform scripts.

Note

Make sure you don't overwrite your existing kxi-terraform dir.
Set the following environment variables in your shell to point to the old and new kxi-terraform directories.
```
export OLD_TF_DIR="<old kxi-terraform dir>"
export NEW_TF_DIR="<new kxi-terraform dir>"
```
Copy the client.ovpn and kxi-terraform.env files from your previous terraform directory to the new directory using the following commands.

Note

Ensure the client.ovpn file is located in the kxi-terraform/terraform/<cloud>/ directory.
```
find $OLD_TF_DIR -type f -name kxi-terraform.env -exec cp {} "$NEW_TF_DIR" \;
cd "$OLD_TF_DIR" && sudo find terraform -type f -name "client.ovpn" -exec cp --parents {} "$NEW_TF_DIR" \;
```

Check that the files have been copied successfully.

find $NEW_TF_DIR -type f -name kxi-terraform.env
find $NEW_TF_DIR -type f -name client.ovpn

Ensure that the manage-cluster docker container is not currently running.

source $NEW_TF_DIR/kxi-terraform.env
docker stop $CLOUD-$ENV-manage-cluster

Add the new TF environment variable to the kxi-terraform.env file.

sed -i '/^TF_VAR_enable_ingress_nginx=/a TF_VAR_enable_nginx_community=true' $NEW_TF_DIR/kxi-terraform.env

Installation steps

To install, follow the steps below:

Build the new kxi-terraform docker image.

cd $NEW_TF_DIR
./scripts/build-image.sh

Run the manage-cluster.sh script.
```
./scripts/manage-cluster.sh
```
Run the terraform init command.
```
./scripts/terraform.sh init config
```
Run the terraform plan command.

Note

The output should only show the new controller that's being added.
```
./scripts/terraform.sh plan config -target="helm_release.nginx-community"
```
Run the terraform apply command.
```
./scripts/terraform.sh apply config
```

Post deploy checks

Run the manage-cluster.sh script.
```
./scripts/manage-cluster.sh
```

Verify the nginx-community helm chart has been deployed.

List any helm releases in the nginx-community namespace.

helm ls -n nginx-community

This should produce the following output.

NAME            NAMESPACE       REVISION        UPDATED                                 STATUS          CHART                   APP VERSION
nginx-community nginx-community 1               2026-03-19 17:20:12.728168049 +0000 UTC deployed        nginx-ingress-2.4.1     5.3.1

Verify the nginx-community pods are running and healthy.

List all pods in the nginx-community namespace.

kubectl get pods -n nginx-community

This should produce the following output.

NAME                                             READY   STATUS    RESTARTS   AGE
nginx-community-nginx-ingress-controller-4xwsm   2/2     Running   0          49m
nginx-community-nginx-ingress-controller-9wrnf   2/2     Running   0          37m
nginx-community-nginx-ingress-controller-ct52n   2/2     Running   0          63m

Note

There should be one pod per Kubernetes node as the controller deployment is a DaemonSet.

Verify the nginx-community controller has an associated service with an external IP or load balancer.

Check for service endpoints in the ingress-nginx namespace.

kubectl get svc -n nginx-community nginx-community-nginx-ingress-controller

This should produce the following output.

NAME                                       TYPE           CLUSTER-IP      EXTERNAL-IP                                                                     PORT(S)                      AGE
nginx-community-nginx-ingress-controller   LoadBalancer   10.218.45.152   k8s-nginxcom-nginxcom-512b7f8bf9-d94f30f006a91773.elb.eu-west-1.amazonaws.com   80:31173/TCP,443:30097/TCP   20h

Verify the nginx-community ingress class is available in the cluster.

Check for the ingress class.

kubectl get ingressclass -n nginx-community nginx-community

This should produce the following output.

NAME              CONTROLLER                     PARAMETERS   AGE
nginx-community   nginx.org/ingress-controller   <none>       10m

Next steps

Once the new ingress controller has been installed you can proceed to Insights Migration Steps.

Removing ingress-nginx

Once kdb Insights Enterprise has been switched over fully to the new nginx-community controller, you can remove the original ingress-nginx controller from your cluster, as follows:

Ensure that the manage-cluster docker container is not currently running.

source $NEW_TF_DIR/kxi-terraform.env
docker stop $CLOUD-$ENV-manage-cluster

Set the Terraform Variable to ingress-nginx to false.

sed -i -e "s/TF_VAR_enable_ingress_nginx=.*/TF_VAR_enable_ingress_nginx=false/g" $NEW_TF_DIR/kxi-terraform.env

Run the manage-cluster.sh script.
```
./scripts/manage-cluster.sh
```
Run the terraform init command.
```
./scripts/terraform.sh init config
```
Run the terraform plan command.

Note

The output should show the new controller being added.
```
./scripts/terraform.sh plan config -target="helm_release.ingress-nginx"
```
Run the terraform apply command.
```
./scripts/terraform.sh apply config
```

Verify the ingress-nginx helm chart has been removed.

List any helm releases in the ingress-nginx namespace.

helm ls -n ingress-nginx

This should produce the following output:

NAME            NAMESPACE       REVISION        UPDATED                                 STATUS          CHART                   APP VERSION

Verify the ingress-nginx pods and services have been deleted.

Check for running pods in the ingress-nginx namespace.
```
kubectl get pods -n ingress-nginx
```
This should produce the following output:
```
No resources found in ingress-nginx namespace
```
Check for service endpoints in the ingress-nginx namespace.
```
kubectl get svc -n ingress-nginx
```
This should produce the following output.
```
No resources found in ingress-nginx namespace
```