refactor: name every environment after the hostname it serves

[Mtze]

Stacked on #120.

An environment's directory, its GitHub Environment and its landing host are now the same string:

environments/test1.eduide.student.k8s.aet.cit.tum.de/
environments/staging.eduide.student.k8s.aet.cit.tum.de/
environments/e2e.eduide.student.k8s.aet.cit.tum.de/
environments/eduide.artemis.cit.tum.de/
environments/bonn.eduide.aet.cit.tum.de/

Why

deploy.yml already used one value for both the manifest directory and the GitHub Environment, so those two could never disagree. But both could disagree with the host actually served - and did. tum-production was configured for artemis.aet.cit.tum.de while its wildcard certificate and its DNS both said artemis.cit.tum.de, and nothing tied the name to the host closely enough for anyone to notice.

There is now no mapping to keep in sync, because there is no mapping.

What moved

• Eight directories under environments/.
• The Deploy (dispatch) and Rollback dropdowns.
• deploy-e2e.yml and deploy-staging.yml, which named their environment inline.
• validate.yml and test-deploy-logic.sh, which read the chart version from environments/test1/env.yaml.
• README, AGENTS.md and four docs pages. The landing-host table in environments.md is gone: the directory listing is that table now.

What deliberately did not move

• Namespaces stay short (eduide-test1). A Kubernetes namespace cannot contain dots; spec.namespace states each one.
• Cluster environments keep cluster-<name>. They belong to a cluster, not to a hostname.
• Cluster manifests keep their names - clusters/tum-production.yaml is a cluster, and only the environment of that name was renamed. Prose about the cluster was left alone.

GitHub Environments

Recreated under the new names with their reviewer teams and secrets, then the old ones deleted, along with the pre-2.0.0 theia-prod and theia-staging which no workflow referenced and which still held live secrets.

Secret values cannot be read back, so THEIA_KEYCLOAK_COOKIE_SECRET was regenerated for every environment rather than copied. Harmless here since nothing is deployed yet, but it is not a like-for-like migration.

bonn.… and mannheim.… were created with reviewers and no secrets, matching their previous state - there is still no kubeconfig for the eduide cluster.

Cost, accepted deliberately

/deploy test2.eduide.student.k8s.aet.cit.tum.de on a PR, and eight long options in the dropdowns.

Verified

test-deploy-logic.sh ALL PASS (including the hostname-comment check added in #120), check-agents-md.sh, actionlint, shellcheck -S error, every env.yaml validates against its schema, all relative links resolve, and both a test and the production environment still render through the published 2.1.0 chart from their new paths.

Noted, not fixed

deploy-e2e.yml calls EduIDE/theia-scale-tests/.github/workflows/functional-tests.yml with an environment input, but that workflow has no workflow_call trigger and declares no inputs. That call is already broken and predates this change.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 078a8b6f-9105-4d5b-a335-4a6316eecefa

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}