spellsource-server-deployment

Production delivery

Production server releases are delivered by CI and GitOps rather than by running a local cluster-restart task.

1. Bump the semantic version in the root build.gradle for a server release.
2. Commit the release and push it to develop.
3. The Build Server workflow runs the server test suite and publishes the versioned server image.
4. Image automation detects the new semantic tag and reconciles the production Helm release.

Monitor the CI workflow and the approved deployment dashboard until the new release is healthy. Do not document or commit credentials, cluster configuration, or private GitOps locations here.

Preview Helm Chart

helm install --dry-run --debug --generate-name ./src/main/helm/spellsource

Testing Locally (macOS)

1. Install dependencies.

   brew install docker minikube helm

2. Create an ingress tunnel.

   minikube start --vm=true # or minikube start --vm=true --driver=hyperkit
   minikube addons enable ingress registry

3. Create a tunnel.

   minikube tunnel

4. Retrieve the service IPs.

   kubectl get svc

Creating an EKS cluster:

eksctl create cluster --name hiddenswitch-cluster-1 --version 1.19 --fargate

Using images from ghcr.io

Create a personal access token with packages:read permissions. Then create a manifest for the secret and reference it:

---
apiVersion: v1
kind: Secret
metadata:
  name: image-pull-secret
stringData:
  .dockerconfigjson: '{"auths":{"ghcr.io":{"auth":"yourusername:ghp_generatedtoken"}}}'

Then, in your values.yaml:

imagePullSecrets:
 - name: image-pull-secret