Why this exists
We already fixed the Helm chart so people can use any ingress they like (#2932). That part is done.
What is left is our own setup for automated checks and local smoke tests that spin up a small Kubernetes cluster. That setup still installs Ingress NGINX, which the Kubernetes project has stopped maintaining. Staying on it is not a good long-term plan for us or for anyone copying our scripts.
What we want to do
Move the smoke / CI path to Gateway API — the newer, supported way Kubernetes steers people for routing traffic into a cluster. We will pick a maintained Gateway implementation that works well in kind (our tiny test clusters) and update the scripts and docs so the story matches.
What success looks like
- New contributors can run the same “try it locally” flow without depending on retired software.
- Our CI keeps proving that the chart still deploys and responds over HTTP, but through Gateway API instead of the old ingress controller we used before.
Related
- Follow-up to #2932 for test and CI setup, not the chart’s ingress fields themselves.
- Background from the Kubernetes project: Ingress NGINX retirement.
Why this exists
We already fixed the Helm chart so people can use any ingress they like (#2932). That part is done.
What is left is our own setup for automated checks and local smoke tests that spin up a small Kubernetes cluster. That setup still installs Ingress NGINX, which the Kubernetes project has stopped maintaining. Staying on it is not a good long-term plan for us or for anyone copying our scripts.
What we want to do
Move the smoke / CI path to Gateway API — the newer, supported way Kubernetes steers people for routing traffic into a cluster. We will pick a maintained Gateway implementation that works well in kind (our tiny test clusters) and update the scripts and docs so the story matches.
What success looks like
Related