Use a k8s cached client for the branch planner implementation

Open luizbafilho opened this issue 2 years ago • 1 comments

We should use a Kubernetes cached client to optimize terraform objects query since that could potentially put a lot of pressure on k8s API.

This a follow-up issue of #779 derived by Michael comment: https://github.com/weaveworks/tf-controller/pull/790#discussion_r1270837915

User Story

As a Terraform Controller user, I'd like to use a Kubernetes cached client for Terraform objects query in the branch planner implementation, so that it reduces the pressure on the k8s API and optimizes performance.

Acceptance Criteria

[ ] Implement a Kubernetes cached client for querying Terraform objects
[ ] Verify reduced pressure on the k8s API

Jul 21 '23 16:07 luizbafilho

Rewriting as a (controller-runtime) controller with a manager will solve this by default (see #722 for discussion).

Nov 02 '23 11:11 squaremo