What happened?

pod_scheduling_durating_seconds is recording the time that a Pod is gated.

We use the timestamp when the scheduler inserts the pod into the queue:

What did you expect to happen?

The period of time when a Pod fails PreEnque (like being gated) shouldn't be accounted in the pod_scheduling_duration_seconds.

How can we reproduce it (as minimally and precisely as possible)?

Create a Pod with scheduling gates.
Wait some time before removing the gate.
Observe the pod_scheduling_duration_seconds metric

Anything else we need to know?

No response

Kubernetes version

1.26+

Cloud provider

OS version

# On Linux:
$ cat /etc/os-release
# paste output here
$ uname -a
# paste output here

# On Windows:
C:\> wmic os get Caption, Version, BuildNumber, OSArchitecture
# paste output here

Install tools

Container runtime (CRI) and version (if applicable)

Related plugins (CNI, CSI, ...) and versions (if applicable)