Ciao everyone,
I’m working on OSIRIS JSON, a vendor-neutral specification for describing infrastructure resources and their relationships as portable point-in-time snapshots.
OSIRIS does not require SaaS middleware, external AI agents, MCP layers, or third-party services to run. After several years across IT and OT infrastructure projects, I started drafting this specification with the goal of generating a local, portable JSON document from existing infrastructure sources, making end-to-end infrastructure easier to understand in an open, automated, and repeatable way.
To validate the idea in real-world scenarios, I already experimented by building producers in Go for Cisco, AWS, and Azure. I am now exploring the development of a Google Cloud producer openly with the developer community, and I would really appreciate guidance from experienced Google Cloud and Go developers.
The goal is simple: run the OSIRIS JSON producer on-premise manually or as a scheduled job, read Google Cloud infrastructure using Google Cloud SDK libraries, and emit an OSIRIS JSON document.
That document could then be used for documentation, topology diagrams, audits, configuration drift analysis, CMDB/IPAM/DCIM workflows, or controlled AI/context workflows without giving those tools direct access to the Google Cloud environment.
At this early stage, I would especially appreciate feedback on which Google Cloud objects should be mapped first, for example projects, folders, VPC networks, subnets, firewall rules, routes, Compute Engine instances, GKE clusters, Cloud Run services, load balancers, Cloud Storage resources, IAM relationships, and dependencies between services.
This is not intended to replace Google Cloud tooling or become a control plane. It would simply be a read-only external producer that generates a vendor-neutral snapshot.
I would really appreciate advice, feedback, or collaboration from people who automate, document, or operate Google Cloud environments.
Ciao from Italy,
Tia