Grafana Skill
The open observability platform.
Ecosystem Graph
graph LR
grafana["Grafana"]
grafana -- "integrates with" --> prometheus
grafana -- "integrates with" --> loki
grafana -- "alternative to" --> datadog
Quick Start
Grafana connects to disparate data sources (Prometheus, Loki, Postgres) and visualizes them on highly customizable dashboards.
docker run -d -p 3000:3000 grafana/grafana
Production Patterns
Provisioning as Code
Do not click around the Grafana UI to create critical production dashboards. Use Grafana's File Provisioning system (JSON/YAML) or Infrastructure as Code (Terraform) to version control your dashboards and alerts.
Architecture & Scaling
Unified Alerting
Grafana handles alerting natively. Define alert rules directly on the dashboard panels. Ensure Grafana is connected to a Notification Channel like Slack, PagerDuty, or Discord to route critical anomalies immediately.
Error Recovery
If Grafana loses connection to its underlying database (SQLite by default, but should be Postgres in production), the UI will hang. Ensure high availability by clustering Grafana nodes backed by a managed PostgreSQL database.
Security Notes
Never expose Grafana to the public internet without enforcing SSO (Single Sign-On) via Google, GitHub, or Okta. Disable basic authentication (admin/admin) immediately upon provisioning.
Relationships
Alternatives: datadog
Works Well With: prometheus, loki
References
Why use this skill
Use this when your agent works with grafana — structured patterns beat pasted docs and prevent common hallucinations.
AI pitfalls
- Using outdated SDK or API versions from training data
- Inventing environment variable names
- Omitting error handling and retry logic
Production checklist
- Secrets in environment variables, not source code
- Error handling and logging in place
- Rate limits and timeouts configured
Related skills
prometheus— integrates withloki— integrates withdatadog— alternative to
Last Verified: 2026-07-02