Helm repo

Helm charts are a perfect way to deploy your apps to kubernetes. The downside is that you require a mechanism to prevent you from duplicating charts projects and keeping them up-to-date.

This project streamlines the way you manage your charts in a central location and reuse them in all of your projects. We make use of the s3 plugin to store the charts in a managed solution at a very low cost.

Usage

The goal is to create one (or a few) repositories that contain multiple charts. By using this project in your pipeline, you can publish the charts to s3.

The project will look at any charts found in /charts in the container and publish them to the s3 bucket defined in the SC_BUCKET environment variable.

docker run --rm \
    -v $(pwd)/charts:/charts:ro \
    -e S3_BUCKET=s3://com.example.helm/charts \
    stenicbv/helm-s3-repo

AWS credentials

Because you don't want your s3 bucket to be open to the world, you will have to inject your AWS credentials into the container. You have the option to inject the ~/.aws folder together with the AWS_PROFILE you want to select or by injecting your AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY. If you are running the container in an AWS environment, you can also make use of an instance profile to handle authentication.

Using your AWS_PROFILE

docker run --rm \
    -v $(pwd)/charts:/charts:ro \
    -e S3_BUCKET=s3://com.example.helm/charts \
    -v ~/.aws:/root/.aws \
    -e AWS_PROFILE \
    -e AWS_DEFAULT_REGION \
    stenicbv/helm-s3-repo

Using credentials

docker run --rm \
    -v $(pwd)/charts:/charts:ro \
    -e S3_BUCKET=s3://com.example.helm/charts \
    -v ~/.aws:/root/.aws \
    -e AWS_ACCESS_KEY_ID \
    -e AWS_SECRET_ACCESS_KEY \
    -e AWS_DEFAULT_REGION \
    stenicbv/helm-s3-repo

Getting started

Bucket

To get the project up and running, you will need to create an s3 bucket to push your charts to. The following is a terraform snippet to create your bucket.

resource "aws_s3_bucket" "helm-repo" {
   bucket = "com.example.helm"
   acl    = "private"
}

Repository

Let's setup a repository containing some example charts we want to publish. The location of the charts is not as important, as long as the charts are all in the same directory.

charts
├── web-application
└── streams-application

Charts pipeline

Now we create a script to publish the charts in the directory. We mount the location where the charts are to /charts in the container and define the S3_BUCKET.

docker run --rm \
    -v $(pwd)/charts:/charts:ro \
    -e S3_BUCKET=s3://com.example.helm/charts \
    stenicbv/helm-s3-repo

Projects

Now that our charts have been published we can start using them. The implementation in the projects will be reduced to a minimum.

deploy/
├── my-project
│   ├── Chart.yaml
│   └── values.yaml

Chart.yaml

apiVersion: v2
name: my-project
description: A Helm chart for Kubernetes
type: application
version: 0.1.0
appVersion: 0.1.0
dependencies:
- name: web-application
  version: "0.1.0"
  repository: "s3://com.example.helm/charts"
  alias: app

Note alias setting. This will set a common prefix to the configuration.

values.yaml

app:
  image:
    repository: nginx

Deploying a project

As an addition step during the deployment of your app, you will need to make sure the chart dependency is resolved before you can install the charts.

helm dependencies update

This command will fetch the chart from the central repo.

Configuration

The project provides a limited set of configuration options.

Parameter	Description	Default
S3_BUCKET	The bucket to push the charts to. s3://com.example.helm/charts	(Required)
S3_BUCKET_ACL	The ACL to apply to any uploaded artifacts	public-read