Description
What happened?
What happened?
After restarting an ephemeral Grafana pod (either using the Helm chart or an external Grafana instance), the Grafana OnCall plugin does not automatically reconnect. Instead, the OnCall page displays the following error message:
"Plugin not connected. Looks like OnCall plugin hasn't been connected yet or has been misconfigured. Retry or contact your administrator."
To fix this issue, we must manually run the following API command inside the Grafana container
curl 'http://admin:admin@127.0.0.1/grafana/api/plugins/grafana-oncall-app/resources/plugin/install' -X 'POST'
Without this manual intervention, the plugin remains in a broken state.
What did you expect to happen?
The Grafana OnCall plugin should automatically reconnect and function normally after a pod restart without requiring a manual API request.
Did this work before?
Unknown. We encountered the issue while deploying Grafana v11.5.2 with Grafana OnCall v1.15.2 on Kubernetes v1.30.9-eks-8cce635.
How do we reproduce it?
How do we reproduce it?
- Deploy Grafana v11.5.2 with Grafana OnCall v1.15.2 on Kubernetes v1.30.9-eks-8cce635.
- Restart the Grafana pod (via Helm upgrade, pod deletion, or another method).
- Navigate to the Grafana OnCall plugin page.
- Observe the "Plugin not connected" error.
- Run the following command inside the Grafana pod:
curl 'http://admin:admin@127.0.0.1/grafana/api/plugins/grafana-oncall-app/resources/plugin/install' -X 'POST'
Refresh the Grafana OnCall page; the plugin now loads correctly.
I've also raised #5490 but this feels like it could be more of a grafana bug.
Is the bug inside a dashboard panel?
No response
Environment (with versions)?
Environment (with versions)?
- Grafana: v11.5.2
- OS: Kubernetes (EKS) v1.30.9-eks-8cce635
- Browser: Google Chrome v112.0.5615.137,
Grafana platform?
None
Datasource(s)?
No response