Skip to main content

How to Deploy

This guide outlines the deployment command sequences to release the tupynambalucas.dev ecosystem across local and production environments using Skaffold and Kubernetes.

1. Kubernetes Environments (Skaffold)​

Deployments across all bounded contexts (platform, cortex, studio) are managed via Skaffold, orchestrating the build, tag, and deploy phases using Kustomize manifests.

1.1. One-Time Deployments​

To compile all workspace images and apply their Kubernetes manifests directly to the active cluster context without hot-reloading:

pnpm infra:up

This builds the containers using the local engine cache, hydrates the image tags inside the manifests, and applies them to the cluster.

1.2. Continuous Development (Live Reload)​

For active cluster testing with live logs and automatic redeploys when source code changes:

pnpm infra:dev

1.3. Terminating Cluster Deployments​

To delete all deployed resources and namespace allocations from the cluster:

pnpm infra:clean

2. Production Deployments​

For production environments, we use Kustomize overlays located in infrastructure/environments/prod.

  1. Ensure your kubectl context points to your production cluster.
  2. Apply the production overlays directly (or via a GitOps operator like ArgoCD/Flux):
kubectl apply -k infrastructure/environments/prod/platform
kubectl apply -k infrastructure/environments/prod/cortex
kubectl apply -k infrastructure/environments/prod/studio

Note: For production, images should be pre-built and pushed to a remote registry (e.g., via GitHub Actions) before applying the manifests, rather than building them locally with Skaffold.