Is there a way to start an application, in a docker container, during deployment to Kubernetes, using Helm?

Question

I am trying to avoid having to create three different images for separate deployment environments.

Some context on our current ci/cd pipeline:

For the CI portion, we build our app into a docker container and then submit that container to a security scan. Once the security scan is successful, the container gets put into a private container repository.

For the CD portion, using helm charts, we pull the container from the repository and then deploy to a company managed Kubernetes cluster.

There was an ask and the solution was to use a piece of software in the container. And for some reason (I'm the devops person and not the software engineer) the software needs environment variables (specific to the deployment environment) passed to it when it starts. How would we be able to start and pass environment variables to this software at deployment?

I could just create three different images with the environment variables but I feel like that is an anti-pattern. It takes away from the flexibility of having one image that can be deployed to different environments.

Can any one point me to resources that can accomplish starting an application with specific environment variables using Helm? I've looked but did not find a solution or anything that pointed me to the right direction. As a plan b, I'll just create three different images but I want to make sure that there is not a better way.

Answer 1

Depending on the container orchestration, you can pass the env in differnt ways:

Plain docker:

docker run -e MY_VAR=MY_VAL <image>

Docker compose:

version: '3'
services:
  app:
    image: '<image>'
    environment:
      - MY_VAR=my-value

Check on docker-compose docs

Kubernetes:

apiVersion: v1
kind: Pod
metadata:
  name: envar-demo
spec:
  containers:
  - name: app
    image: <image>
    env:
    - name: MY_VAR
      value: "my value"

Check on kubernetes-docu

Helm:

Add the values in your values.yaml:

myKey: myValue

Then reference it in your helm template:

apiVersion: v1
kind: Pod
metadata:
  name: envar-demo
spec:
  containers:
  - name: app
    image: <image>
    env:
    - name: MY_VAR
      value: {{ .Values.myKey }}

Check out the helm docs.