OS images not found for workload cluster during upgrade

This looks to be a transient issue but a weird one

Seems os-images-info configmap was not refreshed for the upgrade-workload-cluster stage

2025/08/07 19:39:04.392953 HelmRelease/cluster state changed: Progressing - Running 'upgrade' action with timeout of 5m0s
2025/08/07 19:39:04.469932 Kustomization/multus state changed: DependencyNotReady - dependency 'kubeadm-capm3-virt/cluster-machines-ready' is not ready
2025/08/07 19:39:04.596066 HelmRelease/cluster state changed: UpgradeFailed - Helm upgrade failed for release kubeadm-capm3-virt/cluster with chart sylva-capi-cluster@0.9.18: execution error at (sylva-capi-cluster/templates/resources.yaml:1:4): .capm3/.control_plane.capm3: no OS image could be found matching selectors
selectors:
  - hardened: "false"
    k8s-flavor: kubeadm
    k8s-version: 1.31.10
    os: ubuntu
    os-release: noble
OS images:
  ubuntu-jammy-plain-kubeadm-1-30-11:
    _os_images_key_: ubuntu-jammy-plain-kubeadm-1-30-11
    archive-size: "1436121021"
    commit-id: 5b4b1aff12ab490d6799647783cacc4e23a0eec7
    commit-tag: 0.3.13
    compression: gz
    cosign_publickey: |
      -----BEGIN PUBLIC KEY-----
      MFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAEcuuLszwy0u7y394KY6GW1BgjfaU6
      aK7e95MV+ikZnRfR5EHtqEk1tScNhMqQJMpuFs3QH1TVBr6TIWVLx5cUtg==
      -----END PUBLIC KEY-----
    filename: ubuntu-jammy-plain-kubeadm-1-30-11.raw
    flavor: plain
    hardened: "false"
    image-format: raw
    k8s-flavor: kubeadm
    k8s-version: 1.30.11
    md5: 6e96cfa48e241a43d504a1be6a29049c
    os: ubuntu
    os-release: jammy
    sha256: 92d6609ba1a5f1d15bfbf330519655698c427477d1f5f2e80674c1

Job #10952733731 failed for 24bd02c8:

cc @tmmorin @rletrocquer @mihai.zaharia @tmmorin

---

EDIT(@tmmorin): what happens is that there is a possibly race where cluster unit reconciles before os-images-info unit has updated the os-images-info configmap