Goal

• Each HyperShift commit involving the HyperShift Operator gets delivered without manual intervention all the way to ROSA/HCP staging environment

Why is this important?

• Reduce HyperShift engineering release duty toil
• Ensure predictable delivery cadence
• Improve release visibility and tracking with Inscope
• Ensure consistent pipeline stages
• Leverage automated soaking time and roll out

Scenarios

1.  Merge queue in main:
   1. Prow gates that tests HyperShift in the HyperShift CI cluster
      on-merge Konflux build
   2. crt-redhat-acm HyperShift Operator application's Release plan admission pushes to prod-service registry
   3. HyperShift's app-interface saas.yaml gets triggered and modifies the HYPERSHIFT_OPERATOR_OVERRIDING_DATA variable for osd-fleet-manager-integration to the new prod-service build
   4. Trigger ROSA/HCP SD tests example from OSD
   5. Trigger HCP QE regression tests. Currently we trigger manually in Jenkins although there was bash tool
      1. On success, promote to stage
      2. Trigger ROSA/HCP SD tests
      3. Trigger HCP QE regression tests
         1. On success, promote to prod? (maybe not everytime, probably Mon/Wed)

Acceptance Criteria

• Automated roll outs to production
• Troubleshooting documented in SOP

Dependencies (internal and external)

1. app-interface
2. Konflux
3. Inscope
4. OpenShift prow CI

Previous Work (Optional):

1. Kickstart meeting notes

Useful links:

How to set up saas to reference new image

Open questions:

1. What should be the follow-up for stage to prod?
2. Should we use increasing release numbers for versions? Should we use dates? (semver doesn't seem to be in konflux automation ATM)

Done Checklist

• CI - CI is running, tests are automated and merged.
• HyperShift SOP describing the pipeline with troubleshooting steps <link to Feature Enablement Presentation>
• QE - Automated tests merged: <link or reference to automated tests>