Pipeline cleanup scripts suck

**Describe the task**
Our cleanup scripts intermittently fail to cleanup all resources. This results in resources slowly bleeding away until someone goes and does manual cleanups. It also results in errors in various parts of openshift as things get partially deleted.

**Acceptance Criteria**
- [ ] Cleanup always cleans up! Or lets us know when it fails.
- [ ] All relevant build configs are cleaned up. (e.g. wps-ubuntu-pr is not currently being cleaned)

**Additional context**
- Expecting PR imagestream etc. to be deleted. (Only imagestream left should be the one that's copied into the production imagestream)
- List build configs: oc -n e1e498-tools get BuildConfig --no-headers

Start by listing the build configurations with `oc -n e1e498-tools get BuildConfig --no-headers` and trace the pipeline cleanup scripts that remove those resources. Done means cleanup reliably removes all relevant build configs and PR resources, leaves only the production imagestream copy, and reports failures instead of silently leaving partial deletions.

Written by the indexing model from the issue text.