Grafana
grafana · Observability
GitHub stars
77,575Awaiting first sync
No measured star history yet
What is Grafana?
Grafana is a place to look at the numbers behind your systems, whichever database or monitoring tool holds them. Teams connect their existing data sources, build dashboards, set up alerts and share what they find, while the data stays where it already lives.
Grafana OSS is the open-source edition of Grafana. It queries metrics, logs and traces wherever they are stored, shows them on dashboards your team can share, and sends alert notifications when a condition is met. You run it yourself with Docker or an operating system package.
Core capabilities
- Built-in data sources include Prometheus, Loki, PostgreSQL, MySQL, Elasticsearch and Tempo.
- Alert rules send notifications to Slack, email, PagerDuty, Microsoft Teams, webhooks and more.
- Self-host with Docker or Linux, Windows and macOS packages under the AGPL-3.0 licence.
One dashboard for many data sources
Dashboards are made of panels, each showing a visualization of a query. Built-in data sources cover Prometheus, Loki, Tempo, Graphite, InfluxDB, Elasticsearch, PostgreSQL, MySQL, Microsoft SQL Server and the main cloud monitoring services. Panels in one dashboard can use different sources, and template variables turn a dashboard into a reusable view.
- Dashboards
- Prometheus
- Loki
- PostgreSQL
Alerts and sharing for the whole team
Alert rules run on queries across your data sources, and contact points deliver notifications to Slack, email, PagerDuty, Opsgenie, Microsoft Teams, Telegram, webhooks and others. Dashboards can be shared by link, as snapshots kept on your own instance, as an iframe panel embed or as exported JSON. PDF export and scheduled reports need Grafana Enterprise.
- Alerting
- Contact points
- Snapshots
- JSON export
Run it yourself under AGPL-3.0
Start the grafana/grafana Docker image on port 3000, or install from Debian, Ubuntu, RHEL, Fedora, SUSE, Windows or macOS packages, or deploy with the Helm chart. The default SQLite database suits evaluation; for production the docs point to MySQL or PostgreSQL. The code is AGPL-3.0, with some libraries under Apache-2.0. Enterprise features and Grafana Cloud are separate.
- Self-hosted
- Docker
- AGPL-3.0
- PostgreSQL
Where it fits
Use cases
- 01
Build a service health dashboard
Add Prometheus as a data source, then create a dashboard with graph panels for the numbers you watch, such as response times and error rates. Add a template variable so the same dashboard works for each environment or server.
- 02
Get alerted when something breaks
Write an alert rule on a metric or log query, set a threshold, and choose a contact point such as Slack, email or PagerDuty. Grafana evaluates the rule continuously and notifies your team when it fires.
- 03
Investigate an incident across metrics and logs
Open Explore, run a query against a metrics source and switch to a log source such as Loki with your label filters kept. Compare time ranges side by side, then share a link to the findings with your team.