Grafana App Installation

Kure Monitor ships with a dedicated Grafana App plugin (kuremonitor-kure-app) that brings the proactive AI troubleshooting chat and failure feed directly into your existing observability stack.

Because Kure Monitor is a specialized tool, the plugin is distributed independently rather than through the built-in Grafana catalog.

Installation

1. Declarative Installation (Recommended)

If you are deploying Grafana via the official Helm chart or Kubernetes manifests, you can configure Grafana to automatically install the plugin on startup using environment variables.

Add the following to your Grafana values.yaml:

env:
  GF_INSTALL_PLUGINS: "https://github.com/igor-koricanac/kuremonitor-kure-app/releases/download/v1.0.2/kuremonitor-kure-app-1.0.2.zip;kuremonitor-kure-app"
  GF_PLUGINS_ALLOW_LOADING_UNSIGNED_PLUGINS: "kuremonitor-kure-app"

When Grafana boots, it will automatically download the plugin and allow it to load.

2. Manual Installation (CLI)

If you are running Grafana natively or want to test the plugin before committing to your Helm chart, you can install it manually using the grafana cli tool.

Run the following command in your server terminal, or execute it inside your running container/pod:

grafana cli plugins install kuremonitor-kure-app --pluginUrl https://github.com/igor-koricanac/kuremonitor-kure-app/releases/download/v1.0.2/kuremonitor-kure-app-1.0.2.zip

Allowing Unsigned Plugins If you install manually, you must still explicitly tell Grafana to allow the unsigned plugin. Add the following to your grafana.ini configuration file:

[plugins]
allow_loading_unsigned_plugins = kuremonitor-kure-app

(Alternatively, you can set the GF_PLUGINS_ALLOW_LOADING_UNSIGNED_PLUGINS="kuremonitor-kure-app" environment variable).

Post-Installation

1. Restart Grafana to load the new unsigned plugin.
2. In the Grafana UI, navigate to Apps → Kure Monitor and click Enable.
3. Go to the plugin Configuration tab to connect it to your Kure backend:
   • Backend URL: If Grafana runs in the same cluster, use the internal service URL: http://kure-backend.kure-system.svc.cluster.local:8000.
   • API Key (Service Token): You must provide the X-Service-Token. You can retrieve this by reading the bootstrap secret generated during Kure installation:

     kubectl get secret kure-monitor-bootstrap -n kure-system -o jsonpath='{.data.service-token}' | base64 -d

4. Click Save & Test. You will now see the Kure Monitor dashboard, AI Advice, and Troubleshooting Feed fully integrated!