Custom resource status, conditions do not update during long running delete. #836
Description
Is your feature request related to a problem?
Custom resource status, conditions do not update during long running delete.
For example, ElastiCache replication group delete now has requeue logic in place for long running delete.
This results in kubectl describe to return the resource being deleted as output. While delete is still in progress, the output contains following metadata fields:
deletionGracePeriodSecondsdeletionTimestamp
indicating that a delete is in progress.
However, the output also contains:
"status": {
...
"conditions": [
{
"status": "True",
"type": "ACK.ResourceSynced"
}
],
...
"status": "available"
...
}
among other stale ko.status fields.
This feature request is to be able to get the latest and updated conditions information for the resource which is getting deleted.
Describe the solution you'd like
Patch the custom resource with the latest response from service API during long running delete (when requeue is performed during reconciler run) and update the conditions to indicate delete progress.
Describe alternatives you've considered
Reset the ko.status to nil to indicate delete of the resource. But as the latest resource information is available during reconciler run, the other approach which provides delete progress information and updated conditions seems more useful.