'Continuous Delivery Dashboard' shows bundles in not ready state

We are using Rancher 2.6.4 and have switched to a repo with kustomize support.

In the CD dashboard we see that the bundles are 0/0. Also on the Git Repo page we see that the bundles are not assigned correctly, because only the resources are displayed but not the bundles. In Advanced/Bundles page we see that the bundle exist and is active.

Unfortunately it happens sometimes that on the Git Repo page in the tab Conditions a sync error with Git is shown and then Fleet deletes all resources from the cluster. After a force update with git the resources are deployed again.

Now I have 2 questions:

1. Is this with the bundles only a display error?
2. Has the undeployment to do with the missing assignment to the bundle?

Since we are currently only deploying Grafana dashboards with Fleet it is not that critical, but when the application is rolled out from Fleet it would be fatal if an undeployment happens when the connection to Git is broken.

I see that an error has occurred but I have no idea how to better isolate it …

  Conditions:                                                                                                                                                                                                                                                                                            
     Last Update Time:  2022-11-25T20:39:05Z 
     Status:            True
     Type:              Deployed
     Last Update Time:  2022-11-26T06:38:51Z
     Message:           configmap.v1 ipcs-monlog-system/grafana-dashboard-elasticsearch missing
     Reason:            Error
     Status:            False
     Type:              Ready
     Last Update Time:  2022-11-26T06:38:51Z
     Status:            True
     Type:              Monitored

Last change in git was 4 days ago (2022-11-22).

After force reload of the git repo, the resources are deployed as shown in the picture.

Below is our fleet.yaml file

namespace: ipcs-monlog-system
targetCustomizations:
  - name: lab
    clusterSelector:
      matchLabels:
        management.cattle.io/cluster-display-name: labcluser
    kustomize:
      dir: overlays/lab

  - name: prod
    clusterSelector:
      matchLabels:
        management.cattle.io/cluster-display-name: prodcluster
    kustomize:
      dir: overlays/prod

There seems to be a bug in fleet.