Send the production-deploy dispatch to happo/zoo

[trotzig]

happo-view-kubernetes, the Kubernetes manifests and deploy tooling, was imported into happo/zoo as projects/kubernetes (HAP-1141, happo/zoo#116). Its auto-deploy workflow is now zoo's .github/workflows/auto-deploy.yml. The old repository is frozen and will be archived, so a dispatch sent there deploys nothing.

This points publish-docker's Trigger production deploy step at happo/zoo. The event type (docker-image-published) and the payload ({ image, tag }) are unchanged. Nothing else in the repo mentions happo-view-kubernetes.

happo-automations got the same change in happo/happo-automations#193. This is interim: the docs import into zoo will replace the dispatch with a same-repo workflow_call.

The token

HAPPO_VIEW_KUBERNETES_DEPLOY_TOKEN keeps its name. It lives in CircleCI: the docker context, which publish-docker uses, or the project's environment variables.

It needs Contents: Read and write on happo/zoo: a fine-grained PAT with zoo added, or a classic PAT with repo scope. Without that, the dispatch fails with "Resource not accessible by personal access token".

When to merge

Once zoo's deploy environment has KUBECONFIG and AUTO_DEPLOY_GIT_TOKEN set. Until then, a release's dispatch would start a deploy that cannot run.

Testing

prettier --check .circleci/config.yml passes. The dispatch itself only runs on a v* tag, so the next release is its first real run.

🤖 Generated with Claude Code

[trotzig]

I have updated the HAPPO_VIEW_KUBERNETES_DEPLOY_TOKEN