Support of other automation frameworks

[melezhik]

By convention, Ansible is used for provisioning (community.proxmox upstream); Cinc (Chef) handles modular, recursive desired state complexity.

Would you consider other provision tools support? I am author of Raku based automation framework - https://sparrowhub.io - Sparrow could be used instead of Ansible and Cinc

We can discuss here

[melezhik]

Features

• all management tasks could be accomplished via UI or git integration (gitops style you have now) by using Sparky - https://github.com/melezhik/sparky

• there are a plenty of ready plugins to manage linux boxes - https://sparrowhub.io/search?q=all

• Ruby is natively supported as well

• You have it already planned at version 1.1-3 - " tasks as integrated abstraction for automated script execution" - but tasks concept is already first class citizen in Sparrow core, could be probably good fit for this ...

[stevius10]

Hi @melezhik,

Thanks a lot for reaching out and your suggestion! Looks like a very impressive project—I’ll definitely have a closer look, regardless of this context. Love seeing other automation tools :)

I genuinely appreciate you taking the time to think about how it could fit into Proxmox-GitOps. However, integrating another provisioning tool would be a significant architectural change at this stage. The project aims a specific design choice: a self-contained bootstrap process without external dependencies that could be validated in the same context meanwhile.

Here’s a bit more on the current approach: The project uses a set of custom Ruby libraries for infrastructure management that are intended to run both locally and in a plain Ruby context by passing the ctx object (/config/libraries). The core principle is recursive self-containment, where the control plane deploys and validates itself using its own internal logic. This means the libraries are inherently tested (validate, at least in a significant scope) during deployment run, initiating a self-validating mechanism.

While an integration seems challenging right now without huge refactoring, I guess it could be a highly valuable approach to adopting the concept to it.

Thanks for your input!