we need to resolve some compatability issues between the upstream e2e core karpenter tests and how we setup hypershift to run autonode.

this story documents the changes we need to make either upstream or downstream to resolve these issues. From the spike https://issues.redhat.com/browse/PODAUTO-323 I've concluded two main issues:

• The default timeout in the tests is too low, we should bump it up in the upstream
• The upstream tests expect to be creating and re-creating NodeClasses before every test in a suite, but hypershift's karpenter-operator prevents deletions.