What problem does it solve? When a Grafana alert fires, operators need to quickly determine why it triggered, what it affects, and what to do next. This Skill automates that investigation by fetching alert state, querying the underlying metrics and logs, and producing a concise root-cause analysis with actionable next steps. ## Core Features & Use Cases - Alert State Triage: Fetches alert rules via gcx, distinguishes recording rules and healthy inactive alerts from genuinely firing or pending alerts, and stops early when no action is needed. - Datasource Investigation: Queries Prometheus metrics and Loki logs using the alert's datasourceUID, rendering ASCII graph visualizations to reveal spikes, trends, and anomalies. - Actionable Analysis: Extracts runbook and dashboard links from annotations, suggests follow-up queries by error class (connection errors, latency, error rates, resource exhaustion), and recommends incident creation when there is customer impact. - Use Case: An on-call engineer sees a HighErrorRate alert firing in production. The Skill retrieves the alert, graphs the 5xx error rate over the last hour, correlates with Loki error logs, identifies a recent deployment as the likely cause, and lists the next queries to run. ## Quick Start Ask the assistant to investigate why the HighErrorRate Grafana alert is firing and show the current trend and likely causes.
