Check out the docs at https://github.com/actions/actions-runner-controller/blob/master/docs/about-arc.md
Create a bucket for terraform state, making sure you disable public access and allow your account to access it. (or reuse existing that is noted in the environment file)
Create a Github App in your account and install it in the repo you want to provide runners for. All is explained in : https://github.com/actions/actions-runner-controller/blob/master/docs/authenticating-to-the-github-api.md
In your Google Cloud Project create the secrets for
environment_name.env in the folder environments with the following contents:project_id = "PROJECT_ID" # google PROJECT_ID that you want to deploy in
region = "gcp_region" # GCP region for the network
zone = "europe-west3-c" # GCP zone for the nodes
main_runner = {
name = "main-runner" # Main runner pool name
machine_type = "e2-standard-16" # Main runner pool machine type
min_node_count = "1" # Main runner pool minimal node count
max_node_count = "5" # Main runner pool maximal node count
min_replicas = "5" # Min number of runner PODs in the main pool . Do not confuse with Nodes
max_replicas = "20" # Max number of runner PODs in the main pool . Do not confuse with Nodes
webhook_scaling # Enable webhook scaling for main pool
}
environment = "environment_name" # Name of the environment. Used as a prefix like dev- stag- anything-
ingress_domain = "fqdn" # FQDN for webhook ingress
organization = "org" # Github Organization to use runners in
repository = "repo" # Repository to use runners in
github_app_id_secret_name = "app_id_secret_name" # Google secret name for app id
github_app_install_id_secret_name = "install_id_secret_name" # Google secret name for install_id
github_private_key_secret_name = "pem_file_secret_name" # Google secret name for pem file
deploy_webhook = "false" # Terraform to deploy the scaling webhook
#state_bucket_name = "state_bucket_name" # Not used by terraform. This is just to reference what bucket is used for others
If you want to create additonal pools you can use the additional_runner_pools which is a list of objects. Example:
additional_runner_pools = [
{
name = "test-runner" # Pool name
machine_type = "e2-standard-2" # Macihne type for the pool
min_node_count = 1 # Minimal node count
max_node_count = 2 # Maximal node count
min_replicas = 1 # Minimal replica count
min_replicas = 2 # Maximal replica count
webhook_scaling = true # Enable webhook based scaling
runner_image = "gcr.io/someimage:sometag" # Image to use
labels = ["self-hosted", "testrunner"] # Label set for runner pool. Used in `on`
enable_selector = "true" # Enables NodeSelector, forcing runners to this pool
enable_taint = "true" # Enables Taints. Prevents other runner pods to run in this pool.
requests = { # K8s cpu and memory requests
cpu = "500m" #
memory = "500mi"} #
limits = { # K8s cpu and memory limits
cpu = "2" #
memory = "2Gi"}}] #
Make sure you set the bucket name in the comment in the environment file for documentation purposes
From this directory, login to your gcloud account that you created the bucket with and init terraform with:
gcloud auth login gcloud auth application-default login terraform init -backend-config="bucket=bucket_name"
terraform apply -var-file=environments/environment_name.env
If you need to update the configuration (e.g. upgrading the github runner image, changing resource settings, etc), follow the steps below:
beam-arc-state.gcloud auth login gcloud auth application-default login terraform init -backend-config="bucket=bucket_name"
beam.env. Fix config problems if any.terraform plan -var-file=environments/environment_name.env
beam.env.terraform apply -var-file=environments/environment_name.env
get_kubeconfig_command terraform output and run the command