Grafana
Also known as: Grafana Labs, Grafana dashboard, LGTM stack
Grafana is a visualization platform for operational data.
You connect it to Prometheus, databases, cloud services, or any of 150+ data sources, and build dashboards with graphs, gauges, tables, and heatmaps. It's the standard tool for DevOps and platform teams to monitor system health and business metrics.
Grafana, explained at three levels
At home
Grafana turns raw numbers and data into visual dashboards. Connect it to where your data lives, and Grafana draws graphs, gauges, and tables. It's like a car's instrument cluster -- all the important readings in one visible place.
At work
Teams use Grafana dashboards to monitor their systems at a glance. Operations teams see server health; product teams track user activity; management sees business KPIs. All the data comes from different sources but appears in one interface.
At home
Grafana panels (graph, bar chart, table, heatmap, stat, gauge) are configured with queries against a connected data source. Dashboards can have variables (dropdowns to filter by environment, region, or service). Alerts can fire when panel values cross thresholds, routing to email, Slack, or PagerDuty.
At work
Grafana supports 150+ data sources: Prometheus, InfluxDB, Elasticsearch, PostgreSQL, MySQL, BigQuery, CloudWatch, Azure Monitor, and more. Loki adds log aggregation (like Elasticsearch but optimized for Prometheus-style label-based querying). Tempo adds distributed trace visualization. Together they form Grafana's LGTM stack (Loki, Grafana, Tempo, Mimir).
At home
Grafana dashboards are stored as JSON, making them version-controllable and deployable as code. Grafonnet (a Jsonnet library) and Terraform's Grafana provider allow dashboard-as-code workflows. Provisioning via YAML allows automatic setup of data sources and dashboards on startup.
At work
Grafana Enterprise adds fine-grained RBAC, team-level dashboard permissions, reporting (PDF/email), caching, and enterprise data source plugins. Grafana OnCall integrates on-call scheduling and incident management. At large scale, Grafana's data source query federation and cross-data-source correlations (linking metrics to logs to traces) provide the full observability workflow in one tool.
Real-world examples
At home
Aisha connected Grafana to her home server's Prometheus instance and built a dashboard showing CPU, memory, and disk usage across all her machines. She set an alert to email her if disk usage exceeds 90%.
At work
Carlos's team has a Grafana dashboard with three rows: top row shows infrastructure health (Prometheus), middle row shows application errors (Loki), and bottom row shows business KPIs (PostgreSQL). Every incident review starts with this dashboard.
Common confusions
- Grafana stores your data — Grafana is a visualization layer -- it queries data from your existing sources (Prometheus, databases, etc.) but stores nothing itself. Your data stays where it lives; Grafana just reads and displays it.
- Grafana is only for DevOps teams — Any team with data in a supported data source can use Grafana. Product teams track feature usage, finance tracks spending, sales teams monitor pipeline metrics. It's a general-purpose dashboarding tool.
For analysts
Business analysts use Grafana to build executive dashboards by connecting it to a PostgreSQL or BigQuery data source. Dashboards refresh automatically, and variable dropdowns allow filtering by region, product, or time range without modifying queries.
History & origin
Grafana was created by Torkel Odegaard and released in 2014 as a fork of Kibana (part of the ELK stack), redesigned to work with Graphite and then Prometheus. It grew into a standalone project, became company-backed (Grafana Labs formed in 2014), and expanded from metrics visualization to logs (Loki), traces (Tempo), and a full observability stack. Grafana Cloud launched in 2019 as a managed SaaS version.