Requeue delay is always 1 min regardless of delay time requested

[ewassef]

Describe the bug
When a requeue is returned from the controller, the item is placed on the queue with a 1 minute interval regardless of what the code sends in
To Reproduce

In your ReconcileAsync method simple return return ResourceControllerResult.RequeueEvent(TimeSpan.FromMinutes(15));

Expected behavior

dbug: KubeOps.Operator.Kubernetes.ResourceWatcher[0]
      The server closed the connection. Trying to reconnect.
dbug: KubeOps.Operator.Controller.ManagedResourceController[0]
      Execute/Reconcile event "Reconcile" on resource "ClusterEntity/auto-sample-await".
dbug: ClusterController[0]
      Entered reconcile - state Verified - 4/25/2022 3:31:03 PM
info: KubeOps.Operator.Controller.ManagedResourceController[0]
      Event type "Reconcile" on resource "ClusterEntity/auto-sample-await" successfully reconciled. Requeue requested with delay "00:15:00".
dbug: KubeOps.Operator.Kubernetes.ResourceWatcher[0]
      The server closed the connection. Trying to reconnect.
dbug: KubeOps.Operator.Controller.ManagedResourceController[0]
      Execute/Reconcile event "Reconcile" on resource "ClusterEntity/auto-sample-await".
dbug: ClusterController[0]
      Entered reconcile - state Verified - 4/25/2022 3:32:03 PM <<== This shouldnt be every min, but should be every 15 mins

Additional context
This line seems to have something to do with the issue :
dbug: KubeOps.Operator.Kubernetes.ResourceWatcher[0] The server closed the connection. Trying to reconnect.

[erin-allison]

The watcher has a default timeout interval of 60 seconds, so it will always reconcile each resource no less frequently than once per minute.

Since a reconcile event enters the queue at that interval, the event with 15 min delay is dropped.
I forgot the de-duplication stuff isn't merged yet; the event w/ 15 minute delay should execute in addition to the "every one minute" that comes from the watcher reconnecting.

If you want to increase the timeout interval, adjust OperatorSettings.WatcherHttpTimeout to a higher value. For your example, maybe up to 1800 (30 minutes) so you can see that the re-queue is functioning.

https://github.com/buehler/dotnet-operator-sdk/blob/c90ad7ca78fcdeef7fb5b119061d3404334c34a0/src/KubeOps/Operator/OperatorSettings.cs#L108-L111

[ewassef]

Thank you for the details... ill try that

[ewassef]

Closing as this work-around works