Skip to main content
⏳ Estimated read time: 4 min read

Grafana

Connect Aiden with your Grafana platform to get intelligent insights from your observability stack through a unified interface. Once enabled, Aiden can leverage your Grafana datasources to analyze metrics, logs, and other observability data through natural language conversations.

Integrations Capabilities​

With Grafana integration, Aiden can:

  • Access and analyze metrics from Prometheus datasources.
  • Query and investigate logs from Loki datasources.
  • Compare metric trends across different time periods.
  • Perform detailed root cause analysis (RCA).
  • Analyze firing alerts.
  • Correlate logs and metrics to help troubleshoot issues.
  • Investigate performance bottlenecks and anomalies.
  • Analyze resource utilization patterns.
  • Generate insights from your observability data.

Supported Datasources​

The Grafana integration currently supports accessing the following datasources:

  • Prometheus: For all metrics data.
  • Loki: For all logging data.
  • Read or Manage Grafana based alerts.
  • Read the list of Grafana dashboards.

Additional support for datasources will be added in future updates.

Create Custom Skills​

On top of the several out of the box capabilities, you can create custom skills to add additional capabilities. Here are some examples:

  • Create automated health checks based on Grafana datasources
  • Set up cross-datasource correlation for complex troubleshooting
  • Develop custom anomaly detection for specific services

Check out the Skills page for more information.

Enable Grafana Integration​

Generate your Grafana API Key​

note
  • You need to be an Admin in Grafana.
  • Security Best Practices:
    • Create a dedicated API key for Aiden with minimum required permissions.
    • Use an API key with at least the Editor permission as Aiden will need to query the Datasources which needs the ability to Explore.
    • Regularly rotate API keys (every 60-90 days recommended).
    • Ensure your Grafana API key has appropriate permissions for the datasources you want Aiden to access. For using Prometheus and Loki datasources, the API key needs at least Editor role for these datasources.

  1. Navigate to Adminstration > Users and access > Service accounts.

    service accounts

  2. Click the Add Service Account button and follow the on-screen instructions to create a Service Account.

    add service account

  3. Create a Service Account Token by clicking the Add Service Account Token button and follow the on-screen instructions:

    Generate token

Steps to Enable Grafana Integration​

Follow these steps to enable the Integration:

  1. From the Aiden home page, click Integrations from the navigation panel to the left.

  2. Hover and click the Activate button on the Grafana Integration card.

  3. Enter the integration configuration parameters:

    grafana

    • Grafana URL: This is the web address (endpoint) of your Grafana server. It tells Aiden where to send API requests for dashboards, panels, alerts, and metrics.
      • If Grafana is self-hosted, you can find this in your deployment or ingress configuration.
      • If you’re using Grafana Cloud, the URL will look like: https://<your-stack>.grafana.net.
      • Check out HTTP API Overview.
    • API Token: This is the authentication token Aiden uses to access the Grafana API securely.
      • You can create one by navigating to Grafana > ⚙️ Settings > API Keys > New API Key.
      • Choose an appropriate role and permissions:
        • Viewer: Read-only access (recommended for monitoring).
        • Editor: Modify dashboards.
        • Admin: Full control (use sparingly).

  4. Click Save to enable the integration.

Sample Prompts​

Here are a few sample prompts that you can use:

  • Show me CPU usage trends for the payment service over the last 24 hours.
  • What error logs correlate with the latency spike at 3 PM yesterday?.
  • Compare memory usage between production and staging environments.
  • Are there any services showing abnormal error rates?
  • Find all error logs from the authentication service during the last incident.
  • Which services have the highest request latency right now?
  • Create a report of resource utilization trends for all production services.
  • Tech store orders are failing. Find out why.
  • Analyze the resource usage of pods in prod and provide recommendation for right-sizing.

Additional References​

Last updated on