During the post-pivot restore phase of an IBU upgrade, the LCA polls every 30 seconds (requeueWithShortInterval) to check if a restore has completed. In practice, most individual restores complete in under 5 seconds (confirmed via Velero logs), so the controller spends most of the restore phase idle waiting rather than processing.

* requeueWithShortInterval (30s): https://github.com/openshift-kni/lifecycle-agent/blob/release-4.20/controllers/ibu_controller.go#L103-L105
* Called from HandleRestore: https://github.com/openshift-kni/lifecycle-agent/blob/release-4.20/controllers/upgrade_handlers.go#L550-L552
    

lifecycle-agent v4.20.1 (the value is hardcoded and has not changed across versions)
    

Always. The 30s interval is hardcoded and applies to every IBU upgrade with OADP restores.

    

    1. Configure an IBU upgrade with multiple restore waves
    2. Trigger the upgrade
    3. After the reboot, observe the LCA logs during the restore phase
    4. Note the 30s gap between each restore completion
    

The LCA waits 30 seconds between restore status checks, even when the restore completes in under 5 seconds.
    

The polling interval should be reduced to 5-10 seconds to minimize idle waiting and improve the overall IBU upgrade time.
    

Note that requeueWithShortInterval is also used in other parts of the LCA code.