Deploy Monitoring through Apps - Issue : page not found

[seb-835]

Hi Team,

when i install the Monitoring apps throught Rancher Management UI :
all goes fine, i reach Grafana through Monitoring-> Grafana and i got to the home dashboard.

when i install using the managedCharts from the template

addons:
  monitoring:
    enabled: true
    version: 105.2.0+up66.7.1-rancher.10

Monitoring-> Grafana bring me to the Grafana Page not found :

If i click the "home link" on this page then i get the wanted home dashboard.

I have the same behaviour when installing through helm, so i think there is a missing parameters, but can't find which yet.

[zackbradys]

hey @seb-835, did you verify that version of the monitoring chart is compatible with the version of Rancher? If you go to a cluster's dashboard, then click Cluster, then Tools, then Monitoring, is that the same version?

If you are in an airgapped environment do you have all the required images available? Any pods failing or erroring under the cattle-monitoring-system namespace?

[seb-835]

Hi @zackbradys ,
yes, exactly the same, i take version from from Rancher UI and all images are available and containers are running in cattle-monitoring-system namespace.
Prometheus/Grafana are working well, it s just the first redirection from rancher ui to grafana that link to a not found page... but as soon as i select home or a dashboard, all is fine.

Best.

[seb-835]

Hi, in fact this a problem of chicken and eggs, or making the deployment in 2 pass (downstream then apps).

As for dowstream clusters to avoid the Grafana 404 page not found when reaching through monitoring>grafana, you have to fill the monitoring value file with :

addons:
  monitoring:
    enabled: true
    version: 105.2.0+up66.7.1-rancher.io
    values:
      global:
        cattle:
          clusterId: xxxx
          clusterName: xxxx
      grafana:
        grafana.ini:
          auth.anonymous:
            org_role: Admin
        sidecar:
          dashboards:
            searchNamespace: ALL
      rke2ControllerManager:
        enabled: true
      rke2Etcd:
        enabled: true
      rke2IngressNginx: 
        enabled: true
      rke2Proxy:
        enabled: true
      rke2Scheduler:
        enabled: true

but clusterId is only available after downstream cluster deployment is done and through a human query :

kubectl get clusters.management.cattle.io -o json | jq -r '.items[] | select(.spec.displayName == "my-cluster-0") | .metadata.name'