How to customize NGINX Ingress controller with Helm

byGonzalo Marcos -
In one of our last posts, we saw how to install NGINX Ingress controller using a Helm chart. Helm is a great tool to manage and standardize our solutions deployed in K8s. It gives us great control and consistency to deploy the same configurations across our environments and clusters.

As we discussed in this post, there are two ways of customizing a Helm chart - using the —set flag in the command line or modifying the values.yaml file. In general, the first one is what we’d do if we only needed to change very few parameters of the installation. The second option is a preferred approach for installations that require many changes or changes that span multiple settings.

Let’s see a practical example with NGINX. For us, Mulesoft Architects and Developers, it’s very important for our Runtime Fabric setups to understand how to adapt the NGINX ingress controller to our needs. We’ll see in this post both approaches: how to apply simple and quick changes with the —set flag and how to use the values.yaml file. Let’s see it:

Prerequisites

To follow these examples, we’ll need to have


Initial setup

Create a dedicated namespace

  • We’ll create a dedicated namespace for all the ingress resources called ingress-nginx
kubectl create ns ingress-nginx


Add the Helm repository

  • For this example we will use the Bitnami repository. To add run the command
helm repo add bitnami https://charts.bitnami.com/bitnami
  • Verify that the new repo is in the list of repos of your Helm installation
helm repo list
  • Update the new repo
helm repo update


Explore the default values. 

  • To understand what can be customized, it's always useful to inspect the default values of the NGINX chart:
helm show values bitnami/nginx
  • This will output the default values.yaml file, which lists all the configurable options.


Basic customizations

Let’s say we’re happy with the default values and we just want to customize the following:
  • The number of replicas to run the controller (it would be number of replicas in the deployment). We’ll set it up to 3.
  • The type of service - Instead of using a LoadBalancer service we want to use a NodePort service.
  • The port for our NodePort service. We want to use port number 30092 in all worker nodes.
In this case, this can be easily customize with the following command:

helm install gon-nginx bitnami/nginx -n ingress-nginx --set replicaCount=3 --set service.type=NodePort

Where:
  • gon-nginx - it’s the name of our release
  • -n ingress-nginx - specifies the namespace where we want to deploy all the resources of the chart
  • replicaCount and service.type are the parameters we are customizing
We should get an outcome like this:


  • Verify there’s a new release in the namespace with:
helm list -n ingress-nginx
  • Next, verify we’ve got 3 replicas running the ingress controller
kubectl get pods -n ingress-nginx
  • And lastly, let’s verify that the service is of a type NodePort
kubectl get svc -n ingress-nginx


Create our custom values.yaml file

Now let’s do a more complex customization. Based on the default values, create a custom my-values.yaml file to override the settings. Here's the my-values.yaml file for that will use for our example:

replicaCount: 2

image:
repository: bitnami/nginx
pullPolicy: IfNotPresent

service:
type: NodePort

resources:
limits:
cpu: 500m
memory: 256Mi
requests:
cpu: 250m
memory: 128Mi

livenessProbe:
enabled: true
initialDelaySeconds: 30
periodSeconds: 10
timeoutSeconds: 5
failureThreshold: 6
successThreshold: 1

readinessProbe:
enabled: true
initialDelaySeconds: 5
periodSeconds: 10
timeoutSeconds: 3
failureThreshold: 3
successThreshold: 1

persistence:
enabled: true
accessModes:
- ReadWriteOnce
size: 8Gi
path: /usr/share/nginx/html
annotations: {}
storageClass: ""

ingress:
enabled: true
hostname: my-nginx.local
annotations:
kubernetes.io/ingress.class: nginx
tls: []

Where:
  • replicaCount: Specifies the number of NGINX pod replicas. Here, it's set to 5.
  • image: Defines the Docker image to use for NGINX. You can customize the repository, tag, and pull policy.
  • service.type: Set to LoadBalancer to expose the NGINX service externally. The default port is 80.
  • resources: Configures the resource requests and limits for the NGINX pods.
  • livenessProbe and readinessProbe: Enable and configure health checks to ensure NGINX is running correctly.
  • persistence: Enables persistent storage for NGINX. You can specify the size, access mode, and path.
  • ingress: Configures ingress settings to expose NGINX via a hostname and manage TLS if required.


Install the Chart

Once you've created your values.yaml, you can install the NGINX chart using Helm with your custom configuration:

helm install gon-nginx-2 bitnami/nginx -n ingress-nginx -f my-values.yaml


Verify the installation

helm list -n ingress-nginx
kubectl get pods
kubectl get svc

Check the status of the pods, services, and ingress to ensure everything is configured correctly.


And if we get the details of one of the pods we can get the values of the resources and readinessProbe:


Combining both approaches

Let’s just see with one example how to customize a chart using a custom values.yaml file and the command line. Let’s say that now we basically want to do a custom installation like the one we did in the second example, using the same my-values.yaml file but, this time, we want to set the number of replicas to 5 for testing. 

We can then leave the my-values.yaml as it is (and keep it as a standard for our installations) and change the number of replicas with the —set flag.

We could do that running the following command that combines both approaches:

helm install gon-nginx-3 bitnami/nginx -n ingress-nginx -f my-values.yaml --set replicaCount=5


Verify that now the number of replicas is 5


How to upgrade or rollback if neeeded

If you need to make further changes to your configuration, modify the my-values.yaml and upgrade the release:

helm upgrade gon-nginx bitnami/nginx -f my-values.yaml

If something goes wrong, you can rollback to the previous version:

helm rollback gon-nginx 1

Tags
Previous Post Next Post