Grafana Alerts Agent

This skill provides a workflow for auditing active alerts and maintaining an up-to-date ALERTS.md report.

Workflow

1. Fetch Grafana Alerts

Use the list_alert_rules tool to retrieve all Grafana-managed alert rules. Filter for those in a firing or pending state.

2. Fetch Alertmanager Alerts

If an Alertmanager datasource is available, use list_alert_rules with the datasourceUid to fetch Prometheus/Loki managed alerts.

• •Tip: Use list_datasources(type='prometheus') or list_datasources(type='loki') to find relevant UIDs if unknown.

3. Analyze Alert Severity and Impact

For each active alert:

• •Identify the service or component affected.
• •Determine the severity (critical, warning, info).
• •Extract relevant labels and annotations (summary, description).

4. Update ALERTS.md

Create or update ALERTS.md in the project root with a structured report of active alerts.

• •Format the report with clear headings, tables, or lists.
• •Include timestamps for when the alert started.

5. Report to Main Thread

Provide a concise summary of the active alerts to the user, highlighting any critical issues and confirming that ALERTS.md has been updated.