EPIC Goal

What are we trying to solve here?

Tech debt issues

• Today we have a tonne of boilerplate tekton code in https://gitlab.cee.redhat.com/rhidp/rhdh/-/tree/rhdh-1-rhel-9/.tekton but it could be smarter if we move to reuse snippets with ttps://github.com/konflux-ci/olm-operator-konflux-sample/tree/main/.tekton

• we also use a pre-gen gitlab pipeline step in https://gitlab.cee.redhat.com/rhidp/rhdh/-/blob/rhdh-1-rhel-9/build/ci/sync-midstream.sh (not sure how to remove this, unless we move it upstream into the GH repos?)

• and we have lots of old .in files from the CPaaS days, which could be scrubbed and scrapped to make the repo cleaner and simpler (done as of 2025/06/26)

Why is this important?

Simplification and deduping of boilerplate? Oh, obvious reasons.

Acceptance Criteria

instead of a GL repo with 3 child folders and some shared infra, we would have individual GH repos and could use git submodules + a GH-to-GL sync (like with the RHDH docs repo) to drive Konflux.

If we can move the transformation code into Containerfiles then we might not even need the midstream anymore.

That said we apparently DO have to use GL because:

• we need ELS images for FIPS (RHIDP-1100)
• we need Internal Konflux for ELS images
• therefore we need to use GL as the input to the builds