Skip to content

Question about Pull Mode Architecture and Permissions #652

@kahirokunn

Description

@kahirokunn

Context

I'm reviewing the Pull Mode documentation and trying to understand the architecture and permission model.

Questions

1. Architecture Comparison with Work API

Is the Pull Mode architecture the same as the Work API from the Kubernetes multicluster SIG? If there are differences, could you clarify what they are?

2. Permission Model Clarification

Based on my understanding, the management cluster:

  • Has read permissions to managed clusters (required for features like EventFramework)
  • Does not have write permissions to managed clusters

Is this understanding correct? Can Sveltos be configured to ensure the management cluster has no write access to managed clusters?

3. Zero-Access Configuration

If EventFramework is not used for a particular cluster, is it possible to configure Sveltos so that the management cluster has zero access (neither read nor write) to the managed cluster? In other words, can it operate purely as a Work controller where the managed cluster pulls all necessary information?

Thank you for your help in clarifying these architectural and permission aspects! 🙇‍♂️

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions