Deploy to Kubernetes Action

This action will run the kubectl commands to deploy an image to Kubernetes, and create all of the relevant release tracking information in Github's Deployments.

Inputs

Input	Required	Default	Description
token	yes		Github Token (to create the Deployment record). Typically provide ${{ secrets.GITHUB_TOKEN }}.
namespace	no		Namespace of the Kubernetes Deployment
deployment	yes		Deployment name
container	yes		Container name within the deployment
image	yes		Image to set in the container
environment	yes		Github environment name
ref	no	Current commit hash	ref to attach to the deployment
production	no	(varies)	Boolean indicating if this is an environment users will interact with
transient	no	(varies)	Boolean indicating if this is an environment that may go away in the future
url	no		URL for where the deployment is accessible
wait	no	true	Whether to wait for the rollout to complete before finishing
wait-timeout	no	5m	Maximum time to wait for the rollout. If this timeout is reached, the step will fail and the deployment will be marked as failed on GitHub. Ignored if wait is false.

Outputs

Output	Description

Status tracking

Starting in v0.3.0, this action will by default watch the rollout in the cluster and try to keep the status in GitHub in sync. If the rollout does not complete within the timeout window (wait-timeout), the Deployment will be marked as failed. It is advisable to tune the timeout based on your deployment circimstances; any value accepted by the kubectl rollout status's --timeout flag will work (3s, 5m, etc).

Note: long deployments will result in increased billable time. To disable rollout tracking entirely, run this action with wait: false. This will cause the step to finish when the deployment command is run without regard for completion or success (beyond the command failing outright), which was the default behavior in previous versions. If you deploy very frequently, your deployments take a long time, or are cost-sensitive, this may be a better choice for you. See About billing for GitHub Actions for more info.

Authentication

This action uses kubectl internally to deploy images; you must already be authenticated to the cluster (and be using the correct context).

Depending on a number of factors (e.g. managed vs self-run, which provder, etc.), this process can be cluster-specific. If you are deploying to a managed Kubernetes cluster, it's best to follow any guidance they offer on how to securely authenticate within Github Actions.

However, a bare-minimum guide is included in AUTHENTICATION.md.

Example

The following Actions workflow file will:

• Authenticate to the cluster
• Deploy the image

on:
  push:
    branches:
      - main

jobs:
  deploy:
    name: Deploy head of main to prod
    runs-on: ubuntu-latest
    steps:

      - name: Get GKE credentials
        uses: google-github-actions/get-gke-credentials@main
        with:
          cluster_name: ${{ secrets.GKE_CLUSTER_NAME }}
          location: ${{ secrets.GKE_CLUSTER_LOCATION }}
          credentials: ${{ secrets.GCP_SA_KEY }}

      - uses: firehed/deploy-to-kubernetes-action@v0.3.0
        with:
          namespace: github-actions
          deployment: www
          container: server
          image: my/image:${{ github.sha }}
          token: ${{ secrets.GITHUB_TOKEN }}

Known issues/Future features

• You must authenticate to your cluster before this action
• Specifically if using Google's get-gke-credentials action, if you are checking out code during the workflow job, you must do so before running that step (it creates a file, and checking out code after will remove that file)
• Due to some conflicting magic behavior on Github's end, use of this action will produce strange results if used in conjunction with jobs.<job_id>.environment configuration. See #4