What problem does it solve? When an SLO starts breaching or burning error budget, operators need to quickly find the root cause instead of manually piecing together Grafana dashboards, PromQL queries, and alert rules. This Skill runs a structured investigation workflow that pinpoints the failing dimension and surfaces relevant runbooks. ## Core Features & Use Cases - Guided Investigation Workflow: Retrieves the SLO definition, checks status with early exits for OK/NODATA states, renders timelines, and performs dimensional breakdown of ratio or freeform queries. - Alert and Runbook Correlation: Searches alert rules by SLO name or UUID, extracts runbook and dashboard URLs from annotations, and fetches runbook content from GitHub via the gh CLI. - PromQL Reference: Includes a reference of Grafana SLO recording rule metrics (SLI, burn rate, error budget) with ready-to-use query patterns. - Use Case: An on-call engineer receives a fast-burn alert for an availability SLO, asks the agent to investigate, and gets a report showing the worst offending cluster, related firing alert rules, and the runbook's remediation steps. ## Quick Start Ask the agent to investigate why your SLO is breaching by providing its name or UUID, for example: investigate why the checkout-availability SLO is burning error budget.
