Prod/non-prod software artifacts issues with groupped services pattern

[Brightside56]

It's probably similar to #3411

In some pipelines I've ended up with multiple warehouses, where:

1. one warehouse with non-production artifacts can be used for non-production stages (tags with short commit hashes, xxxxxx)
2. another warehouse with production artifacts can be used for production stages (semver tags, 1.2.3)

It's working great for common case pattern https://docs.kargo.io/user-guide/patterns/#common-case

However for grouped services https://docs.kargo.io/user-guide/patterns/#grouped-services it may be much more complicated.

If I have non-prod warehouse (with service-a/xxxxxx, service-b/xxxxxx and service-c/xxxxxx images) and it's production ready equivalent - freight B from prod warehouse (with service-a/1.2.3, service-b/1.2.3 and service-c/1.2.3), then there are situations when team may have on dev environment somewhat like service-a/xxxxxx, service-b/1.2.3, service-c/1.2.3, but there is no possibility to produce such freights while having two warehouses.

Then some team member finds service-b/1.2.3 and service-c/1.2.3 xxxxxx-equivalents to create desired fright in non-prod warehouse, then someone needs to re-verify that and there is complexity to find or keep this "mapping" in mind or all the time search production counterparts for non-production images or vice versa

I could have single warehouse for both production/non-production artifacts, however in case when team produces 10-20 non-prod artifacts for service daily - then another kind of difficulty arises when user have to navigate through a large number of images to find image he/she needs

I'm wondering how this experience could be improved? I'm curious how others are solving similar problems and what I could improve

[Brightside56]

I think it can be closed, because it doesn't seem to be useful to have separate warehouses for prod/non-prod artifacts