Why does a pop-up error message appear stating, "Invalid dashboard UID in the request" when attempting to refresh the data?
Issue description:
In high availability (HA) multi-node setups, Grafana runs with two replicas. However, these replicas do not share a common data source, which leads to issues in managing user sessions. Users may encounter errors such as:
- Dashboard not found
- Invalid dashboard UID in the request
Cause:
Lack of a shared data source between Grafana replicas in the HA configuration.
Affected Versions:
- 23.4.0 to 23.4.9
- 23.10.0 to 23.10.6
Fixed Versions:
- 23.4.10
- 23.10.7
Workaround Steps:
- Patch the Helm parameter in the fabric-installer ArgoCD application:
kubectl patch application fabric-installer -n argocd --type='json' -p='[{"op": "add", "path": "/spec/source/helm/parameters/-", "value": {"name": "global.rancherMonitoring.grafana.replicas", "value": "1"}}]'
- Sync the rancher-monitoring application from the ArgoCD UI.