Tracking OncePerOrderedCleanup adoption — call for participation

[Daniel Sionov]

Hi all,

Following the discussion in [#13786](https://github.com/kubevirt/kubevirt/pull/13786) and the Ginkgo author's interest in a `SharedResourceGroup` decorator ([ginkgo#1292](https://github.com/onsi/ginkgo/issues/1292)), I've opened a tracking issue to collect CI impact data from PRs adopting `OncePerOrderedCleanup`:

https://github.com/kubevirt/kubevirt/issues/16935

Early results are encouraging — converting the Prometheus Endpoints tests from per-test VM setup to a shared `BeforeAll` cut the block from ~20 min to ~3 min by eliminating 52 redundant VMI boot cycles ([#16914](https://github.com/kubevirt/kubevirt/pull/16914)).

If you have test blocks where specs create identical VMs and only read from them, consider converting them. The pattern is straightforward — add `Ordered` + `decorators.OncePerOrderedCleanup` and move `BeforeEach`/`AfterEach` to `BeforeAll`/`AfterAll`.

The more data we gather, the stronger the case for the upstream `SharedResourceGroup` decorator that would give us the same benefits with randomized execution.

Contributions and feedback welcome.

Thanks,
Daniel