Although there is no longer a panic and the CGU works as expected applying policies in batches, the status is reported differently and in a misleading way. It appears that the CurrentBatchRemediationProgress gets set to nil when the CGU is completed and the conditions now show Completed instead of TimedOut even when the first batch did time out. It also shows a message saying all clusters are compliant with all policies, even if this is untrue.

    showed in TALM from brew versions v4.19.0-38, v4.18.1-14, and v4.16.4-9 in CI

    always

    1. Create CGU with max concurrency of 1 and two clusters where the first cluster is powered off
    2. Wait for the first cluster/batch to time out and the second cluster/batch to succeed.
    3. Check conditions/run oc get on the CGU and notice that it says Completed with message "All clusters are compliant with all the managed policies" even though .status.clusters[0].currentPolicy.status is NonCompliant     

    Condition type Succeeded has reason Completed

    Condition type Succeeded has reason TimedOut

[klaskosk@klaskosk-thinkpadp1gen3 ~]$ KUBECONFIG=~/kniqe16-kubeconfig oc get cgu -n talm-test talm-cgu
NAME       AGE   STATE       DETAILS
talm-cgu   11m   Completed   All clusters are compliant with all the managed policies