fix(errors): retry transient "failed calling webhook" apiserver 500s

[HsiuChuanHsu]

Fixes #14220

Motivation

A webhook can be briefly unreachable during normal operation. While it is unreachable, the kube-apiserver
rejects the request with an HTTP 500 InternalError:

Internal error occurred: failed calling webhook "..."

An earlier attempt (PR #16038) treated all 500s as transient which was not that correct. Blindly retrying them would mask real failures and pile load onto an already-unhealthy apiserver.
So this task is to retry only this known-transient case while leaving every other 500 classified as permanent.

Modifications

Added a single targeted classifier in util/errors/errors.go, same style as the existing isResourceQuotaTimeoutErr:

func isTransientWebhookErr(err error) bool {
    return apierr.IsInternalError(err) && strings.Contains(err.Error(), "failed calling webhook")
}

And added into isTransientErr` alongside the other transient checks.

Verification

Documentation

AI

Claude Sonnet 4.8

• Debug core logic.
• Refine comment wording
• Unit tests