Add infrastructure topology extractor for Kubernetes and cloud resources

[ravisuhag]

Context

Data assets don't exist in isolation — they run on infrastructure. Service dependencies, Kubernetes deployments, and cloud resource relationships are essential context for AI agents investigating issues or planning changes.

Scope

New extractors for infrastructure metadata:

Kubernetes

• Extract deployments, services, configmaps, secrets (metadata only)
• Capture service-to-service dependencies from service mesh or network policies
• Emit depends_on relationships between services and the data assets they consume

Cloud Resources (GCP/AWS)

• Extract resource metadata: compute instances, managed databases, storage buckets, IAM policies
• Capture resource-to-resource dependencies
• Link cloud resources to the data assets they host

Design Considerations

• Read-only access to cluster/cloud APIs
• Support namespace/project filtering to scope extraction
• Focus on relationships and topology, not operational metrics

Why

When an AI agent investigates why a dashboard is showing stale data, knowing that the upstream service was recently redeployed or that the database instance was scaled down is essential operational context.

References

• Meteor Roadmap — Expand Coverage