Grafana is an open-source platform for querying, visualizing, and alerting on data from virtually any source. Connect it to Prometheus, Elasticsearch, OpenSearch, ClickHouse, PostgreSQL, CloudWatch, or any of 150+ supported data sources -- then build interactive dashboards that give your team a single pane of glass across the entire stack.
Created by Torkel Odegaard in 2013 as a fork of Kibana v3 refocused on time-series metrics, Grafana has grown into one of the most widely used visualization tools in the industry. Grafana Labs, the company behind it, was founded in 2014 and now serves over 7,000 organizations, including 70% of the Fortune 50. Over 20 million users rely on Grafana worldwide.
Key Features
Dashboards: Drag-and-drop dashboard builder with template variables, auto-refresh, and dozens of panel types -- time series graphs, gauges, heatmaps, tables, stat panels, histograms, geo maps, and more. Dashboards can be shared via links, snapshots, or exports. Thousands of community-contributed templates are available.
Data Source Plugins: Grafana doesn't store data. It queries it in real time from whatever backends you're already running. Over 150 data source plugins cover time-series databases (Prometheus, InfluxDB, Graphite), log systems (Loki, Elasticsearch, OpenSearch), relational databases (PostgreSQL, MySQL, MSSQL), cloud monitoring services (CloudWatch, Azure Monitor, Google Cloud Monitoring), tracing backends (Tempo, Jaeger, Zipkin), and many others.
Alerting: Define alert rules visually based on query thresholds, anomaly detection, or composite conditions. Route notifications to Slack, email, PagerDuty, webhooks, or any supported channel. Supports multi-dimensional alerting and silencing.
Explore Mode: An ad-hoc querying interface separate from dashboards. Useful for troubleshooting and investigation when you don't yet know what you're looking for.
Annotations: Mark events -- deployments, incidents, configuration changes -- directly on graphs. Correlating metric changes with specific events speeds up root cause analysis.
Provisioning: Define dashboards and data sources as code through configuration files. Version-control your monitoring setup alongside your infrastructure.
The Grafana Ecosystem (LGTM Stack)
Grafana is the visualization layer, but Grafana Labs has built a full open-source observability stack around it -- commonly called LGTM:
- Loki: Log aggregation designed for cost efficiency. Uses label-based indexing instead of full-text indexing, making it dramatically cheaper to run than alternatives at scale. Think "Prometheus, but for logs."
- Tempo: Distributed tracing backend that requires only object storage. Deeply integrated with Grafana for trace-to-log and trace-to-metric correlation.
- Mimir: Horizontally scalable, highly available long-term storage for Prometheus metrics. Handles the scale challenges that standalone Prometheus struggles with.
- Alloy: Grafana's distribution of the OpenTelemetry Collector. A single agent that collects metrics, logs, traces, and profiling data.
- OnCall: Open-source on-call management integrated with Grafana alerting.
- k6: Open-source load testing tool for performance testing APIs and applications.
- Pyroscope: Continuous profiling database for understanding application resource usage at the line-of-code level.
Deployment Options
Grafana OSS (Free, Self-Hosted): The open-source edition under AGPLv3. Full dashboard and visualization capabilities, core data source plugins, alerting, and community support. You manage the infrastructure.
Grafana Enterprise (Paid, Self-Hosted): Adds premium data source plugins (Splunk, Datadog, Oracle, ServiceNow, Snowflake), advanced security features, reporting, and commercial support. Starts at $8 per active user per month.
Grafana Cloud (Managed SaaS): Fully managed by Grafana Labs with hosted Loki, Mimir, Tempo, and Pyroscope backends. Includes a free tier (10k metric series, 50GB logs, 50GB traces, 3 users). Paid tiers scale by usage -- metrics at $6.50-16 per 1,000 billable series, logs at $0.40/GB ingested. 99.5% uptime SLA.
Grafana Labs also offers a Bring Your Own Cloud (BYOC) option that runs the managed stack inside your own AWS or GCP account, and a Federal Cloud offering for U.S. government compliance (FedRAMP High, DoD IL5).
Common Use Cases
- Infrastructure Monitoring: Visualize server, network, Kubernetes, and cloud infrastructure metrics. Prometheus + Grafana is the de facto standard for Kubernetes monitoring.
- Application Performance Monitoring: Distributed tracing across microservices with Tempo or Jaeger, correlated with metrics and logs on a single dashboard.
- Log Analytics: Centralized log search and analysis through Loki or Elasticsearch/OpenSearch backends. Filter, aggregate, and correlate logs with metrics in one place.
- IoT and Industrial Monitoring: Real-time sensor data visualization -- temperature, pressure, device status -- with high-frequency refresh rates and alerting.
- Business Analytics: SQL database backends (PostgreSQL, MySQL, BigQuery) power sales dashboards, KPI tracking, and operational reporting.
- DevOps and SRE: SLO monitoring, incident response workflows, on-call rotation management, and post-incident review -- all within a single toolchain.
Grafana and the Open-Source Observability Stack
One of Grafana's strengths is that it doesn't lock you into a single backend. Organizations commonly pair Grafana with open-source data backends to build cost-effective observability stacks:
Elasticsearch or OpenSearch handle log aggregation and full-text search. They provide powerful querying, real-time indexing, and rich visualization capabilities through their native dashboards -- and Grafana connects to both as first-class data sources.
ClickHouse serves as a metrics and analytics backend where sub-second queries across billions of rows matter. Its compression ratios and query speed make it well-suited for high-cardinality observability data at a fraction of the cost of commercial alternatives.
Prometheus remains the standard for Kubernetes and infrastructure metrics collection, with Grafana as its natural visualization companion.
Building and operating this kind of stack requires expertise across multiple systems -- configuring ingestion pipelines, tuning storage and retention, optimizing query performance, and ensuring reliability at scale. Getting it right means each component does what it's best at, while Grafana ties the picture together.
License
Grafana OSS is licensed under AGPLv3 (GNU Affero General Public License v3), an OSI-approved open-source license. Grafana Labs switched from Apache 2.0 to AGPLv3 in April 2021 to prevent cloud providers from offering Grafana as a managed service without contributing changes back. Plugins, agents, and certain libraries remain Apache-licensed.