fix(helm): make the cluster DNS domain configurable

[shekhargit1912]

The chart hardcoded cluster.local when building in-cluster service URLs, so installs on clusters configured with a different DNS domain rendered unreachable addresses that had to be patched by hand after every upgrade.

Add a clusterDomain value (default cluster.local) and use it to build the A2A base URL, the nginx controller upstream, and the bundled PostgreSQL connection string. Existing installs are unaffected: when the value is unset the templates render exactly as before.

Fixes #1767

[Copilot]

Pull request overview

This PR makes the Kubernetes cluster DNS domain configurable in the Helm chart so in-cluster service URLs don’t assume cluster.local, addressing installs on clusters with a custom cluster-domain (Fixes #1767).

Changes:

• Added a new clusterDomain value (defaulting to cluster.local) in values.yaml.
• Updated helper templates to build the controller A2A base URL and nginx upstream authority using the configured cluster domain.
• Updated the bundled PostgreSQL connection string template and added Helm unit tests covering custom clusterDomain.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
helm/kagent/values.yaml	Introduces clusterDomain value and updates the documented default for controller.a2aBaseUrl.
helm/kagent/tests/ui-deployment_test.yaml	Adds a test asserting UI internal backend URL uses the custom cluster domain.
helm/kagent/tests/controller-deployment_test.yaml	Adds tests asserting A2A base URL and bundled Postgres URL incorporate the custom cluster domain.
helm/kagent/templates/controller-deployment.yaml	Uses clusterDomain when constructing the bundled Postgres connection string.
helm/kagent/templates/_helpers.tpl	Adds kagent.clusterDomain helper and switches URL helpers to use it.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[EItanya]

Rather than adding this, should we just remove the DNS domain completely. Usually just name.namespace will do the trick

[shekhargit1912]

Rather than adding this, should we just remove the DNS domain completely. Usually just name.namespace will do the trick
@EItanya
Good point - short names resolve in-cluster via the resolv.conf search domains, so dropping the suffix would fix #1767 without any new config. Before I switch the PR over: are any of these three URLs (the A2A base URL, the nginx upstream authority, and the Postgres connection string) ever consumed outside the pod's DNS context, or do they all resolve strictly in-cluster? If they're all in-cluster, I'm happy to drop clusterDomain entirely and just use name.namespace.svc. Want me to rework it that way?

[EItanya]

Rather than adding this, should we just remove the DNS domain completely. Usually just name.namespace will do the trick
@EItanya
Good point - short names resolve in-cluster via the resolv.conf search domains, so dropping the suffix would fix #1767 without any new config. Before I switch the PR over: are any of these three URLs (the A2A base URL, the nginx upstream authority, and the Postgres connection string) ever consumed outside the pod's DNS context, or do they all resolve strictly in-cluster? If they're all in-cluster, I'm happy to drop clusterDomain entirely and just use name.namespace.svc. Want me to rework it that way?

AFAIK they're all in cluster, cluster.local marks it that way.

[shekhargit1912]

Rather than adding this, should we just remove the DNS domain completely. Usually just name.namespace will do the trick
@EItanya
Good point - short names resolve in-cluster via the resolv.conf search domains, so dropping the suffix would fix #1767 without any new config. Before I switch the PR over: are any of these three URLs (the A2A base URL, the nginx upstream authority, and the Postgres connection string) ever consumed outside the pod's DNS context, or do they all resolve strictly in-cluster? If they're all in-cluster, I'm happy to drop clusterDomain entirely and just use name.namespace.svc. Want me to rework it that way?

AFAIK they're all in cluster, cluster.local marks it that way.
@EItanya
Thanks for confirming! I'll rework the PR to drop clusterDomain entirely and switch the A2A base URL, nginx upstream, and Postgres host to the name.namespace.svc short form. That fixes the original issue on clusters with custom DNS domains without adding a new value. Will push the update shortl

[shekhargit1912]

AFAIK they're all in cluster, cluster.local marks it that way.

@EItanya
Thanks for confirming! Reworked as suggested - dropped clusterDomain entirely and switched the A2A base URL, nginx upstream, and bundled Postgres URL to the name.namespace.svc short form, so they resolve on any cluster regardless of its DNS domain. Tests updated accordingly