[release-1.7] deploying a Sylva 1.6 workload cluster on Sylva 1.7 fails if monitoring is enabled
(Related to #3026)
Context: What had been tested until now was the operation (modification, upgrade) of a Sylva 1.6 workload cluster on Sylva 1.7, but not the fresh install case.
When installing a Sylva 1.6 workload cluster on Sylva 1.7 fails (if monitoring unit is enabled):
unit Kustomization/monitoring did not became ready after 15m0s
Details on Kustomization/monitoring and related resources:
IDENTIFIER STATUS REASON MESSAGE
Kustomization/wc-69124202-rke2-capo/monitoring InProgress Kustomization generation is 1, but latest observed generation is -1
╰┄╴HelmRelease/wc-69124202-rke2-capo/monitoring InProgress HelmRelease generation is 1, but latest observed generation is -1
├┄╴┬┄┄[Conditions]
┆ ├┄╴Reconciling True Progressing Fulfilling prerequisites
┆ ╰┄╴Ready False ValuesError could not resolve ConfigMap chart values reference 'wc-69124202-rke2-capo/wc-69124202-rke2-capo-monitoring-info' with key 'values.yaml': configmaps "wc-69124202-rke2-capo-monitoring-info" not found
╰┄╴┬┄┄[Events]
├┄╴2026-11-12 19:32:45 Normal HelmChartCreated Created HelmChart/wc-69124202-rke2-capo/wc-69124202-rke2-capo-monitoring with SourceRef 'HelmRepository/wc-69124202-rke2-capo/unit-monitoring'
├┄╴2026-11-12 19:32:46 Normal HelmChartConfigured Configured HelmChart/wc-69124202-rke2-capo/wc-69124202-rke2-capo-monitoring with SourceRef 'HelmRepository/wc-69124202-rke2-capo/unit-monitoring'
╰┄╴2026-11-12 19:45:33 (x11 over 12m47s) Warning ValuesError could not resolve ConfigMap chart values reference 'wc-69124202-rke2-capo/wc-69124202-rke2-capo-monitoring-info' with key 'values.yaml': configmaps "wc-69124202-rke2-capo-monitoring-info" not foundThe reason is the following:
- in Sylva 1.6, in workload cluster context, the
monitoringunit valuesFrom refer to the{{ .Values.cluster.name }}-monitoring-infoConfigMap - in Sylva 1.7 and later, this has changed to
rancher-monitoring-info(!7255 (merged))
The gist of the problem is that the ConfigMap in workload cluster context is produced by a ClusterPolicy in the management context. The inconsistency happens when the 1.6 workload cluster expects a ConfigMap which the 1.7 ClusterPolicy does not generate anymore.