Deleting Package before status is Ready does not properly clean up clients

[rjferguson21]

Environment

Device and OS:
App version:
Kubernetes distro being used:
Other:

Steps to reproduce

1. kubectl apply -f package-a.yaml && sleep 1 && kubectl delete -f package-a.yaml

Expected result

Package should not have any Keycloak clients in cluster.

Actual Result

Package leaves lingering Keycloak client

Visual Proof (screenshots, videos, text, etc)

Severity/Priority

Additional Context

Add any other context or screenshots about the technical debt here.

[mjnagel]

May have a path forward here with #1159 - because of the upstream Pepr changes we could do a state check when the finalizer is triggered and if the Package is currently Pending we could return false to exit the finalizer without removing it. Then when the Package CR reconciles (or fails to) it would hit the finalizer watch again and perform the full finalizer operation.

[slaskawi]

Another potential idea worth exploring is using the PersistentVolumeClaim set as whenDeleted: Delete. Since we use the start-dev command for Keycloak we leverage the H2 storage (file-based) and we might want to delete the entire database when we perform cleanup.

However, we may need a K3d upgrade for this (the feature is turned on in 1.32 and uds deploy k3d-core-slim-dev:0.33.1 deploys 1.30.4)

Please take my comment with a grain of salt as I'm a beginner in UDS Core and I may not foresee some further implications.

[slaskawi]

After chatting with @UnicornChance, I realized I misunderstood this issue. It's about a Keycloak Client being a leftover after installing some different package. Having said that it's a purely management issue, so probably should be solved in Pepr.