How do you ensure that you are achieving the Pulled Automatically GitOps principle?

The GitOps principle “Pulled Automatically” states that an automated agent (operator, controller, reconciler) must continuously watch the declared state stored in the canonical repository and pull any change into the runtime environment without human intervention. Option B exactly describes this model. All other options involve push-based deployments, manual triggers, or repo-design decisions that do not guarantee continuous, automated reconciliation.

A. Pushes changes only when “major updates” occur and relies on an external pipeline; this is push-based, not automatically pulled.
C. Repo topology (single vs. multiple) is orthogonal; having many repos does not in itself ensure automatic pull-based reconciliation.
D. Manual scripts/commands break the requirement for automatic, continuous reconciliation—no operator is pulling states.

1. GitOps Working Group, “GitOps Principles,” Principle 4 (Pulled Automatically), CNCF document v1.0, lines 49-55. https://github.com/open-gitops/documents/blob/main/PRINCIPLES.md

2. Linux Foundation, GitOps Fundamentals (LFS165), Chapter 2 “GitOps Principles,” slide 13 – “Pulled Automatically: an operator continuously pulls desired state from Git.”

3. Weaveworks & CNCF, “The GitOps Maturity Model,” Section 2.1 “Continuous Reconciliation,” pp. 4-5. DOI:10.5281/zenodo.5244874.