Deploy failure on `constraints/storage.uniformBucketLevelAccess`

[grayside]

We currently have deploy failures if the org constraint constraints/storage.uniformBucketLevelAccess is being enforced. This org constraint helps organizations enforce use of the recommended practice https://cloud.google.com/storage/docs/uniform-bucket-level-access.

We should be in compliance with this constraint by adding uniform_bucket_level_access = true to the google_cloud_storage_bucket resource. Doing so does not appear to require any other changes, we weren't relying on non-uniform access.

[glasnt]

This change causes failures in the setup job when trying to create new products:

"message": "Cannot insert legacy ACL for an object when uniform bucket-level access is enabled. Read more at https://cloud.google.com/storage/docs/uniform-bucket-level-access",

Related error: googlecodelabs/feedback#1231

The create_new_product management command creates the product, then associates the product image, then creates testimonials. This job wasn't failing, because after re-trying, the product would exist, and so the command wouldn't attempt to recreate it (allowing for an idempotent setup task).

This repo doesn't test for which images are presented, nor if testimonials exist. So while the following may look like a working deployment, it's subtly broken, as this avocado is in no way sparkly.

[grayside]

My reading of the linked issue is that django-storages has a gcs integration that requires non-uniform access. This is potentially an upstream bug, because https://cloud.google.com/storage/docs/uniform-bucket-level-access recommends uniform access.

Proposed Fix

1. Revert PR feat: use Uniform Bucket Level Access in Cloud Storage #115
2. Flag this org constraint as incompatible

Katie, do you see a path to keeping uniform bucket access with django-storages?

Preventing Regression

Fix integration tests: we currently check for completionTime that the final Cloud Run Job finishes. Instead/additionally we should check job status for failure on each Cloud Run Job, or and an e2e test that the outcome of each job is seen in the app. Will file this as a follow-up issue.

[grayside]

We'e proceeded with the proposed fix. Leaving this open to consider if we want to fix compliance with the constraint at an application level.

[github-actions]

This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 7 days