As a power Kubernetes user providing Kubernetes based paas to internal customers, we are not looking for more GUI or, abstractions over helm.
There are already a lot of solutions out there in this area like helmsman/argocd/helmfile/cross plane helm-provider. And we like kubernetes resources + gitops based automations over any other fancy tools.
Most of the time problems around helm is it's string based templating and lack of type safety. That's why timoni looks a more promising solution in this space. It's lack of packages is the limiting factor.
Another interesting approach is kapp-controller and carvel tooling. Packaging helm charts, OCI images, etc as OCI artifacts to use as a combined source is really interesting. We were considering using kapp-kontroller however our current dependence helm, and some architectural concerns caused us to pass on kapp-controller for now.
As to the question of what could be selling points towards a new "Package Manager" would be,
* Timoni like packages that has templating with type safety.
* A large package pool, or
* Abstraction over helm packages that could add type safety, or better yet automatic or at least semi-automatic conversion of helm charts (One can dream ;))
* Full management through kubernetes API / CRDs
* Multicluster management, or fleet management.
As a power Kubernetes user providing Kubernetes based paas to internal customers, we are not looking for more GUI or, abstractions over helm.
There are already a lot of solutions out there in this area like helmsman/argocd/helmfile/cross plane helm-provider. And we like kubernetes resources + gitops based automations over any other fancy tools.
Most of the time problems around helm is it's string based templating and lack of type safety. That's why timoni looks a more promising solution in this space. It's lack of packages is the limiting factor.
Another interesting approach is kapp-controller and carvel tooling. Packaging helm charts, OCI images, etc as OCI artifacts to use as a combined source is really interesting. We were considering using kapp-kontroller however our current dependence helm, and some architectural concerns caused us to pass on kapp-controller for now.
As to the question of what could be selling points towards a new "Package Manager" would be,
* Timoni like packages that has templating with type safety. * A large package pool, or * Abstraction over helm packages that could add type safety, or better yet automatic or at least semi-automatic conversion of helm charts (One can dream ;)) * Full management through kubernetes API / CRDs * Multicluster management, or fleet management.