Feature: I want the image tag for my Metadata and Pipeline Invocation charts managed by the baseline chart.

[ewilkins-csi]

Description

Currently, even though metadata and pipeline invocation are v2 charts, the image tag (i.e. the aiSSEMBLE version) is set in the downstream project chart. This doesn't match our usual pattern, and isn't a great user experience. We should set the image tag in the base chart for these images.

DOD

• Change the Metadata chart to extend from the base Quarkus chart
• The MDA templates for Metadata and Pipeline Invocation values files should not set image tag
• Metadata and Pipeline Invocation should pull the correct image tag without being set explicitly in the values file
• The v1 chart migration in foundation-upgrade should no longer target Metadata or Pipeline Invocation for updates.
• Write a baton migration to remove the tag if it is set to the current aissemble version

Test Script

1. Create a project from the archetype

   mvn archetype:generate

2. Add a data delivery pipeline with provenance enabled to issue-307-pipeline-models/src/main/resources/pipelines (e.g. SparkPipeline.json)
3. Generate the project and resolve manual actions

   mvn clean install

4. OTS ONLY: Deactivate the upgrade-v2-chart-files-aissemble-version-migration migration and change the Chart.yaml file to point to the local copy of the aiSSEMBLE baseline chart for the following apps:
   • pipeline-invocation
   • metadata
   • policy-decision-point
5. Verify the image tag is not set in the following files:
   • issue-307/deploy/src/main/resources/apps/metadata/values.yaml
   • issue-307/deploy/src/main/resources/apps/pipeline-invocation/values.yaml files
6. Build the project without the build cache and follow the final manual action

   mvn clean install -Dmaven.build.cache.skipCache

7. Deploy the project

   tilt up; tilt down

8. Ensure the policy-decision-point comes up successfully using the local project image
9. Execute the pipeline via the invocation service

   curl -X POST http://localhost:8085/invoke-pipeline/start-spark-operator-job -d '{
       "applicationName": "spark-pipeline",
       "profile": "dev"
   }' -H "Content-Type: application/json"

10. Watch the pipeline logs and wait for successful completion. It may take 15-20s for the driver pod to appear. Kill the log follow once the pipeline is finished.

    kubectl logs -f spark-pipeline-driver

11. Query the metadata service and verify that metadata was created during pipeline execution. The metadata make take a couple tries to appear as the first creation of metadata takes some time.

    curl localhost:8082/metadata