We’ve developed a “standard” helm chart the we publish to an internal repo that our developer teams can all use to deploy their apps to our Kubernetes environments. It comes with sane defaults, but allows for some customizations as needed.

We have some teams that want to deploy 2 different deployments of the same app. Basically, they will be the exact same code, but they would use secrets to target different backends, use different ingress URLs, etc. For example, a time-off request app. One deployment for DeptA is accessed by timeoff.depta.com, and connects with sqldb1(with creds). The other deployment for DeptB is access by timeoff.deptb.com and connects with sqldb2(with another set of creds).

Yes, this very simple example could be accomplished within our CICD system, but might not be scalable.

Reading up on subcharts and dependencies, I’m wondering if that is the way to go. I’m thinking I could create a “multi-deploy” chart that uses my “standard” chart as a dependency. The teams could then define each instance of the chart needed in their values file.

The dependencies/requirements documentation, with aliases and stuff seems to require me to define each of the deployments ahead of time. I’m looking for a more dynamic solution.

any thoughts or suggestions?