Konflux would need to tweak the leader election parameters for the Kueue controller. In short, I need to increase the leaseDuration and renewDeadline since on the Konflux clusters fragmentation takes about 30 seconds and with the current configurations the leader crashes because it fails to renew the lease at the time given to it.
Upstream Kueue already support configuring those, 

update: There are some defaults OCP recommends for operator authors to set for the lease configuration.

taken from https://github.com/openshift/enhancements/blob/0f916a52af1a6fbdab0c5b80ae0e66c7a27efb6a/CONVENTIONS.md#handling-kube-apiserver-disruption 

the recommendation are: LeaseDuration=137s, RenewDealine=107s, RetryPeriod=26s.

Those values will be sufficient for Konflux, so in case those will be configured, customizing the lease parameters by the user won't be needed.

There was also a suggestion to configure `ReleaseOnCancel` if the Kueue controller supports it.

Slack thread for context - https://redhat-internal.slack.com/archives/C084N2C6P9U/p1751457510022409