This story is about devoting time to maintaining our work with mig-agnosticd/agnosticd, over the past few months a few things have drifted that we need to address.

The goal is:

• Use the upstream agnosticd as our source for launching dev + lab clusters, https://github.com/redhat-cop/agnosticd
  • This means we do not want to carry specific patches in our fork, https://github.com/konveyor/agnosticd
  • We will need to examine what patches we have in our fork and determine if they  need are still needed, if they are needed let's get them submitted upstream. Additional info below as AI.
• Keep mig-agnosticd in line with what we want to deploy for lab clusters, meaning let's have mig-agnosticd be 'close' to what the lab deploys, we do not need to deploy the exact instance size, counts, parameters, but we should be pretty close so we can recreate the lab clusters from mig-agnosticd with a small tweak if desired.
• A new branch in konveyor/agnosticd and have one local for our development:
  • Create a branch 'konveyor-dev', which will hold our patches for dev and testing.
  • The 'development' branch should be in lock-step with the upstream agnosticd/development. Our changes are never merged here directly. Any changes merged here will be overwritten by whatever is current in the upstream branch unconditionally.

Different Repos:

konveyor/mig-agnostid/konveyor-dev - wrapper, this one is pulled by RHPDS when deploying DEV environments for variable/environment definitions
redhat-cop/agnosticd/development - this is the main branch, production. Any changes to agnosticd itself, and PROD labs go in here

Porduction branch in upstream agnosticd repository is named "development". ( this is essentially main ).
Only permanent fixes to agnosticd itself or prod versions of our lab should be PR-ed there.