If the co/samples object is deleted, the samples operator sometimes fails to restore the Degraded false condition to the conditions array.  This leads to visual inconsistency and may cause generic operator actions (test cases, user interfaces, etc) to display an error.

Cluster operators are required to report all three conditions with the appropriate true/false/unknown status, although it is allowable to delay slightly (seconds) before filling out conditions it should not be multiple minutes before the condition is set to false.

apiVersion: config.openshift.io/v1
kind: ClusterOperator
metadata:
  creationTimestamp: "2019-09-05T03:00:41Z"
  generation: 1
  name: openshift-samples
  resourceVersion: "48455"
  selfLink: /apis/config.openshift.io/v1/clusteroperators/openshift-samples
  uid: 5897ab71-cf89-11e9-9e6a-126169521af2
spec: {}
status:
  conditions:
  - lastTransitionTime: "2019-09-05T03:00:41Z"
    message: Samples installation successful at 4.2.0-0.ci-2019-09-04-220355
    status: "True"
    type: Available
  - lastTransitionTime: "2019-09-05T03:00:41Z"
    message: Samples installation successful at 4.2.0-0.ci-2019-09-04-220355
    status: "False"
    type: Progressing
  - lastTransitionTime: "2019-09-05T03:07:24Z"
    status: "False"
    type: Degraded

Note the transition times (before this, the condition was absent)

It would be acceptable to set Degraded false at the same time Available and Progressing are set, and set Degrade true when an error is actually detected.