Add 4 tools for getting grafana dashboards

docs/data-sources/builtin-toolsets/.nav.yml

@@ -13,6 +13,7 @@ nav:
   - Datetime: datetime.md
   - Docker: docker.md
   - GitHub: github.md
+  - Grafana Dashboards: grafanadashboards.md
   - Loki: grafanaloki.md
   - Tempo: grafanatempo.md
   - Helm: helm.md

docs/data-sources/builtin-toolsets/grafanadashboards.md

@@ -0,0 +1,88 @@
+# Grafana Dashboards
+
+Connect HolmesGPT to Grafana for dashboard analysis, query extraction, and understanding your monitoring setup. This integration enables investigation of dashboard configurations and extraction of Prometheus queries for deeper analysis.
+
+## Prerequisites
+
+A [Grafana service account token](https://grafana.com/docs/grafana/latest/administration/service-accounts/) with the following permissions:
+
+- Basic role → Viewer
+
+## Configuration
+
+=== "Holmes CLI"
+
+    Add the following to **~/.holmes/config.yaml**. Create the file if it doesn't exist:
+
+    ```yaml
+    toolsets:
+      grafana/dashboards:
+        enabled: true
+        config:
+          api_key: <your grafana service account token>
+          url: <your grafana url>  # e.g. https://acme-corp.grafana.net or http://localhost:3000
+          # Optional: Custom health check endpoint (defaults to api/health)
+          # healthcheck: api/health
+          # Optional: Additional headers for all requests
+          # headers:
+          #   X-Custom-Header: "custom-value"
+    ```
+
+    --8<-- "snippets/toolset_refresh_warning.md"
+
+    To test, run:
+
+    ```bash
+    holmes ask "Show me all dashboards tagged with 'kubernetes'"
+    ```
+
+=== "Robusta Helm Chart"
+
+    ```yaml
+    holmes:
+      toolsets:
+        grafana/dashboards:
+          enabled: true
+          config:
+            api_key: <your grafana API key>
+            url: <your grafana url>  # e.g. https://acme-corp.grafana.net
+            # Optional: Additional headers for all requests
+            # headers:
+            #   X-Custom-Header: "custom-value"
+    ```
+
+## Capabilities
+
+| Tool Name | Description |
+|-----------|-------------|
+| grafana_search_dashboards | Search for dashboards and folders by query, tags, UIDs, or folder locations |
+| grafana_get_dashboard_by_uid | Retrieve complete dashboard JSON including all panels and queries |
+| grafana_get_home_dashboard | Get the home dashboard configuration |
+| grafana_get_dashboard_tags | List all tags used across dashboards for categorization |
+
+## How it Works
+
+### Dashboard Query Extraction
+
+When HolmesGPT retrieves a dashboard, it can extract and analyze Prometheus queries from dashboard panels. This is particularly useful for:
+
+- Understanding what metrics a dashboard monitors
+- Extracting queries for further investigation with the Prometheus toolset
+- Analyzing dashboard time ranges and variable usage
+
+### Example Usage
+
+**Finding dashboards by tag:**
+```bash
+holmes ask "Find all dashboards tagged with 'production' or 'kubernetes'"
+```
+
+**Analyzing a specific dashboard:**
+```bash
+holmes ask "Show me what metrics the 'Node Exporter' dashboard monitors"
+```
+
+**Extracting queries for investigation:**
+```bash
+holmes ask "Get the CPU usage queries from the Kubernetes cluster dashboard and check if any nodes are throttling"
+```

docs/data-sources/builtin-toolsets/index.md

@@ -19,6 +19,7 @@ HolmesGPT includes pre-built integrations for popular monitoring and observabili
 -   [:material-clock:{ .lg .middle } **Datetime**](datetime.md)
 -   [:simple-docker:{ .lg .middle } **Docker**](docker.md)
 -   [:material-github:{ .lg .middle } **GitHub**](github.md)
+-   [:simple-grafana:{ .lg .middle } **Grafana Dashboards**](grafanadashboards.md)
 -   [:material-package:{ .lg .middle } **Helm**](helm.md)
 -   [:material-web:{ .lg .middle } **Internet**](internet.md)
 -   [:simple-apachekafka:{ .lg .middle } **Kafka**](kafka.md)

holmes/core/tools.py

@@ -767,6 +767,16 @@ def _load_llm_instructions(self, jinja_template: str):
             context={"tool_names": tool_names, "config": self.config},
         )
 
+    def _load_llm_instructions_from_file(self, file_dir: str, filename: str) -> None:
+        """Helper method to load LLM instructions from a jinja2 template file.
+
+        Args:
+            file_dir: Directory where the template file is located (typically os.path.dirname(__file__))
+            filename: Name of the jinja2 template file (e.g., "toolset_grafana_dashboard.jinja2")
+        """
+        template_file_path = os.path.abspath(os.path.join(file_dir, filename))
+        self._load_llm_instructions(jinja_template=f"file://{template_file_path}")
+
 
 class YAMLToolset(Toolset):
     tools: List[YAMLTool]  # type: ignore

holmes/plugins/toolsets/grafana/common.py

@@ -8,7 +8,7 @@
 
 class GrafanaConfig(BaseModel):
     """A config that represents one of the Grafana related tools like Loki or Tempo
-    If `grafana_datasource_uid` is set, then it is assume that Holmes will proxy all
+    If `grafana_datasource_uid` is set, then it is assumed that Holmes will proxy all
     requests through grafana. In this case `url` should be the grafana URL.
     If `grafana_datasource_uid` is not set, it is assumed that the `url` is the
     systems' URL