Customization

The Terraform Operator uses the terraform-runner as its terraform runner to execute terraform commands. If you don’t want to use the default terraform-runner, you can build your own.

To make the operator use your terraform runner, the Terraform Operator expects the following environment variables:

DOCKER_REGISTRY=docker.io
TERRAFORM_RUNNER_IMAGE=kubechamp/terraform-runner
TERRAFORM_RUNNER_IMAGE_TAG=0.0.4 ## <- this might be different

The above are the defaults that are passed to the operator. In helm, you can override these values by setting the following:

terraformRunner:
  image:
    registry: docker.io
    repository: kubechamp/terraform-runner
    tag: "0.0.4"

Building Your Runner

The runner of course must be a docker container at the end, the implementation in the container is up to you, however, there are few things to keep in mind.

When Terraform Operator creates Kubernetes jobs with the Terraform Runner, it sets some environment variables on the Terraform Runner container. For a technical view, have a look at this code section

Environment Variable	Default value	Description
TERRAFORM_VERSION	-	The Terraform version to install, its taken from the `spec.terraformVersion` field
OUTPUT_SECRET_NAME	-	The Kubernetes secret to add the Terraform outputs
TERRAFORM_WORKING_DIR	`/tmp/tfmodule`	The Terraform working directory
TERRAFORM_WORKSPACE	`default`	The Terraform workspace to use
TERRAFORM_DESTROY	`false`	Indicates whether to run a Terraform destroy
TERRAFORM_VAR_FILES_PATH	`/tmp/tfvars`	The path where var files will be mounted
POD_NAMESPACE	`metadata.namespace`	The Kubernetes namespace where the job is created

Git SSH

If the the spec.gitSSHKey was provided to authenticate against private git repositories, the path to the ssh key will be /root/.ssh/id_rsa.

You need to add the ssh key ssh-add /root/.ssh/id_rsa