Retry KWOK Stage CRD wait

[shayasoolin]

What type of PR is this?

/kind bug

What this PR does / why we need it:

The scale-test CI hit a transient kubectl wait failure while waiting for the KWOK Stage CRD:

.status.conditions accessor error: <nil> is of the type <nil>, expected []interface{}

This PR keeps using kubectl wait, but runs it in short retry loops up to the existing timeout. That lets the CRD object finish populating status/discovery before setup fails.

Which issue(s) this PR fixes:

Fixes #550

Special notes for your reviewer:

This is a follow-up to #611. The previous PR made stage-fast.yaml mandatory; this one makes the CRD readiness wait tolerate the short window where the CRD exists but .status.conditions is not populated yet.

Does this PR introduce a API change?

NONE

Additional documentation e.g., enhancement proposals, usage docs, etc.:

NONE

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.