Analyze time-series metrics and create PromQL alerting and recording rules.
npx skills add https://github.com/grafana/skills --skill prometheus-grafana-2
Please help me install this Agent Skill. Skill: prometheus Source: https://github.com/grafana/skills/tree/main/skills/grafana-lgtm/prometheus Command: npx skills add https://github.com/grafana/skills --skill prometheus-grafana-2
Provides clear guidance for querying, visualizing, and alerting on time-series metrics so engineers can quickly identify performance regressions and reliability issues without guessing PromQL syntax or architecture patterns.
Use the prometheus skill to craft a PromQL query that calculates the 5 minute error rate for service api-server, propose a recording rule, and suggest an alert expression and notification policy.
High-intent search queries and answers about installing and using this skill.
To calculate the 5 minute error rate in PromQL, use rate functions and instant vector selectors over a 5 minute range vector. You can aggregate grouped per-service error rates by applying sum and by() clauses to your metric selectors.
The best way to precompute expensive PromQL queries is by authoring recording rules. Recording rules evaluate heavy PromQL expressions on a schedule and store the results as new time-series metrics, reducing dashboard loading times and query load.
You set up alerting by authoring alert expressions using PromQL and configuring notification routing. Alert expressions trigger based on metric thresholds, and notifications are routed through Alertmanager or Grafana unified alerting to deliver alerts to the right channels.
Yes, Prometheus remote write integrates with Grafana Cloud Metrics and Mimir for scalable metric storage. This architecture pattern allows you to push locally scraped metrics to a centralized remote endpoint for long-term retention and querying.
You need recording rules when direct dashboard queries are computationally expensive or cause performance regressions. Precomputing complex aggregations and rate functions into persistent metrics prevents query timeouts during high cardinality or high traffic analysis.
openoctaThe quick brown fox jumps over the lazy dog.
1k-offProvision Grafana alert rules, contact points, and notification policies via YAML.
n4vyshMonitor cloud-native systems with Prometheus and Grafana Cloud Metrics.
1k-offStore and query Prometheus and OpenTelemetry metrics with Grafana Mimir.
