Prometheus Monitoring

Query and operate Prometheus via its HTTP API. Writes PromQL for instant and range queries, inspects targets, alerts, and rules, accesses series and label metadata, and runs TSDB admin operations to troubleshoot monitoring infrastructure.

This skill turns Claude into a Prometheus power user. It builds PromQL queries for instant and range data, retrieves target/alert/rule status, explores series and label metadata, and performs TSDB admin tasks like snapshots and tombstone cleanup — all over the /api/v1 HTTP API.

When to use

Use when querying Prometheus metrics, writing PromQL, checking scrape targets and alert rules, debugging missing metrics, or building dashboards and alerts on top of a Prometheus backend.

Examples

Debug a noisy alert

Investigate why an alert keeps firing

Write PromQL to show the rate of HTTP 5xx errors per service over the last hour and identify which target is breaching the alert threshold

Capacity range query

Trend memory usage over time

Build a Prometheus range query for container memory working set bytes per pod over the last 24h and flag pods approaching their limits