Provide management of infinite namespaced feature branch deployments to a Kubernetes cluster via Kustomize. This is designed to read from a GitHub repository containing your manifests, structured as such (example):
.
├── blue/
│ ├── base/
│ │ ├── kustomization.yaml
│ │ ├── deployment.yaml
│ │ ├── service.yaml
│ │ └── etc...yaml
│ └── overlays/ # this folder will be created **and managed** for you
│ ├── development/
│ │ └── kustomization.yaml
│ ├── feature-branch-1/
│ │ └── kustomization.yaml
│ └── feature-branch-2/
│ └── kustomization.yaml
└── red/
├── base/
│ ├── kustomization.yaml
│ ├── deployment.yaml
│ ├── service.yaml
│ └── etc...yaml
└── overlays/ # this folder will be created **and managed** for you
├── development/
│ └── kustomization.yaml
├── feature-branch-1/
│ └── kustomization.yaml
└── feature-branch-2/
└── kustomization.yaml
You only need to tell Kustomize about the files in your base folder, per kustomize, the rest is managed for you.
Within each services’ base folder, create an appropriate kustomization.yaml
---
kind: Kustomization
apiVersion: kustomize.config.k8s.io/v1beta1
resources:
- deployment.yaml
- service.yaml
- etc...yaml
This repository creates a namespace, and ensures that your services play nicely with each other within it.
As you deploy a feature for a given service, all of your other services are "deployed" within that namespace, but pointed to the service running in the default namespace via ExternalName
. This allows integration tests to be run to ensure that your new features play nicely with existing services. If deploy a second service within the same namespace, it will behave as expected, and deploy the image appropriately.
Ingress are created, to allow any visual testing, as a subdomain matching your feature (but transformed to kebab-case, ie feature-branch-1.localhost).
-s, --service
- The service to deploy to your cluster-r, --cluster-repo
- GitHub repository that controls your cluster'-i, --target-image
- Remotely hosted target image-n, --namespace
- Desired namespace, or inferred from $GITHUB_REF-t, --tag
- Image tag, or inferred from $GITHUB_SHA-T, --token
- GitHub access token with repos access, NOT $GITHUB_TOKEN--flux
- a manifest is generated to allow Weave Flux to deploy your cluster--dry-run
- the yaml files are printed to stdout—-built
- when paired with —-dry-run, will build your kustomizations and output the built manifests
Check out the overlays to be created:
docker build -t kustomized_namespaces/create:latest .
docker run -rm kustomized_namespaces/create:latest -r dudo/k8s_colors -n feature_branch_1 -s blue -i dudo/blue -t latest --dry-run
And to see Kustomize do its thing:
docker run -rm kustomized_namespaces/create:latest -r dudo/k8s_colors -n feature_branch_1 -s blue -i dudo/blue -t latest --dry-run --built
Actually commit to your repo!
docker run -rm kustomized_namespaces/create:latest -r dudo/k8s_colors -n feature_branch_1 -s blue -i dudo/blue -t latest -T YourReposDeployKeyToken
on:
pull_request:
types: [labeled]
name: Build, push, and deploy
env:
TARGET_IMAGE: dudo/blue
jobs:
docker:
if: github.event.label.name == 'deploy'
name: Build and push Docker Image
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v1
- name: Registry
env:
DOCKER_PASSWORD: ${{ secrets.DOCKER_PASSWORD }}
DOCKER_USERNAME: ${{ secrets.DOCKER_USERNAME }}
run: docker login -u $DOCKER_USERNAME -p $DOCKER_PASSWORD
- name: Build
run: docker build -t $TARGET_IMAGE:${GITHUB_SHA::7} .
- name: Push
run: docker push $TARGET_IMAGE:${GITHUB_SHA::7}
deploy:
name: Deploy to Cluster
runs-on: ubuntu-latest
needs: docker
env:
SERVICE: blue
CLUSTER_REPO: dudo/k8s_colors
steps:
- name: Kustomized Namespace - Create Overlay
env:
TOKEN: ${{ secrets.TOKEN }}
uses: zyngl/[email protected]
with:
args: --flux