# Metrics Detective

> Query and analyze monitoring data using Grafana, Prometheus, and Loki. Use when investigating performance issues, interpreting dashboards, writing PromQL or LogQL queries, or generating reports from metrics and logs.

- Skill: `lurui1997/metrics-detective` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add lurui1997/metrics-detective`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lurui1997/metrics-detective/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: lurui1997 (https://skillmd.com/u/lurui1997)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/lurui1997/metrics-detective

---


# Grafana

Query, analyze, and report on observability data.

## PromQL Patterns

- Use `rate(metric[5m])` for counters; adjust window to traffic granularity.
- Aggregate with `by (label)` to reduce cardinality noise.
- Use `histogram_quantile(0.99, rate(metric_bucket[5m]))` for latency percentiles.

## LogQL Patterns

- Filter first with label selectors `{app="service"}`, then pipe `|= "error"`.
- Use `| json` or `| pattern` to extract fields for grouping.
- Aggregate log volume with `sum(rate({app="service"}[1m]))`.

## Dashboard Analysis

- Identify the relevant panel by correlating time ranges with incident timestamps.
- Export panel data as CSV for offline spreadsheet analysis when needed.

## Alert Investigation

- Read the alert expression and compare current values against thresholds.
- Check for label matches between firing alerts and affected workloads.

## Automation

- Use the Grafana API (`/api/datasources/proxy/...` or snapshot endpoints) to programmatically fetch query results.
- Return findings as structured Markdown with query, value, and interpretation.