Who Grafana is for#
SRE and on-call teams monitoring production infrastructure
Grafana centralizes the metrics, logs, and traces that on-call engineers need during incidents. Connecting Prometheus for system metrics and Loki for application logs gives a single interface to identify and investigate problems without switching tools mid-incident.
Skip if:
If your team has no existing Prometheus or Loki setup and needs a fully managed, zero-configuration monitoring service, Datadog or New Relic offer faster initial onboarding at higher ongoing cost.
Platform engineering teams building internal observability stacks
Teams standardizing observability across multiple services use Grafana as the shared dashboard layer. Template variables let one dashboard cover all services; alerting rules deploy via Terraform or GitOps workflows. The open plugin architecture supports custom integrations for internal data sources.
Skip if:
Teams without dedicated infrastructure ownership who cannot commit to running Prometheus, Loki, and Grafana may find the operational overhead outweighs the cost savings versus a managed SaaS alternative.
Developers adding dashboards to self-hosted applications
Developers running their own servers for applications, databases, or services use Grafana to build dashboards for system metrics and application-level KPIs. With a Prometheus exporter and a single Grafana instance, you get production-grade dashboards for any process that exposes metrics.
Skip if:
If you need application performance monitoring with automatic code-level tracing, a dedicated APM tool with language agents, such as the Elastic APM stack or OpenTelemetry with a backend, may provide faster time-to-insight.
Data teams visualizing business and operational metrics
Grafana connects to relational databases (Postgres, MySQL) and analytics systems alongside infrastructure data sources, making it usable for business dashboards alongside infrastructure monitoring. Mixed-source panels let you overlay application load with business conversion data in the same view.
Skip if:
For pure business intelligence with non-technical stakeholders who need governed reports, scheduled exports, or embedded analytics in customer-facing products, a purpose-built BI tool like Metabase or Apache Superset is a better fit.
The problem it solves#
Production infrastructure generates three categories of telemetry: metrics, logs, and traces. The problem is that each lives in a different system with a different query language. Metrics land in Prometheus or InfluxDB, logs in Elasticsearch or Loki, traces in Jaeger or Tempo. When something breaks at 3am, the on-call engineer tabs between four dashboards in four different tools, losing context at each jump.
Paid monitoring services address this fragmentation by bundling collection, storage, and visualization into one managed product, but the cost scales with host count and data volume. A team running 50 servers can easily spend several thousand dollars a month on a commercial monitoring platform. For many teams, especially those already running their own infrastructure, that per-host pricing is the core pain: they are paying for a dashboard layer on top of data they already own.
How it solves it#
Unified dashboards across metrics, logs, and traces
Query Prometheus metrics, Loki logs, and Tempo or Jaeger traces from the same dashboard panel. Grafana preserves label context when you pivot between signal types, so clicking a spike on a metric panel can surface the correlated log stream without rebuilding the query. Over 100 official data source plugins are included.
Alerting with multi-channel notifications
Define alert rules visually against any connected data source. When thresholds fire, Grafana routes notifications to Slack, PagerDuty, VictorOps, OpsGenie, or any webhook target. Alert rules are stored as code and can be version-controlled alongside your infrastructure configuration.
Dynamic dashboards with template variables
Parameterize dashboards with dropdown variables that filter all panels simultaneously. A single dashboard can cover all your services when a service variable is populated from Prometheus labels. Reusable dashboard templates mean you build once and apply across every environment.
Mixed data sources in a single panel
Query different data sources in the same graph on a per-query basis. You can overlay a Prometheus metric with a Postgres query result in one panel, which is useful for correlating system load with business data. This works for custom data source plugins as well.
Explore mode for ad-hoc investigation
Switch to Explore mode to run ad-hoc queries against any connected data source without modifying saved dashboards. Split-view lets you compare two queries side-by-side or examine the same data across two time windows. Logs can be streamed live for real-time tail-like behavior.
Strengths and trade-offs#
Strengths
- 76,000+ GitHub stars and active development since 2013Grafana is one of the most starred observability projects on GitHub, with contributions from thousands of developers and backing from Grafana Labs. The repository receives daily commits and has over 14,000 forks, indicating a large ecosystem of plugins and derivative projects. Common integration problems are well-documented and solved across community forums and GitHub discussions.
- No vendor lock-in on data storageUnlike Datadog and New Relic, which require sending your telemetry to their cloud, Grafana works with data sources you already operate. Your metrics stay in your Prometheus instance, your logs in your Loki cluster, and Grafana is the display layer only. You can swap Grafana out without losing any data.
- AGPL-3.0 with self-hosting from a single Docker commandThe open source core is AGPL-3.0 licensed and free to run on your own hardware. Getting a basic Grafana instance running requires one Docker command and no external dependencies beyond a compatible data source. For teams with existing Prometheus or Loki setups, dashboard coverage can begin within minutes of first launch.
- Over 100 official data source integrationsOfficial plugins connect Grafana to Prometheus, Loki, Elasticsearch, InfluxDB, Tempo, MySQL, Postgres, CloudWatch, Azure Monitor, Google Cloud Monitoring, and dozens more. The plugin architecture means community-built integrations can add support for any system with a queryable API.
Trade-offs
- -AGPL-3.0 requires source disclosure for network servicesGrafana's AGPL-3.0 license requires that if you run a modified version as a network service, you must make your modifications available. For teams using the unmodified open source build internally, this is not a concern. For teams embedding Grafana into a commercial product they ship to customers, it requires either releasing those changes or purchasing a commercial license from Grafana Labs.
- -Grafana is a visualization layer, not a data collectorGrafana does not collect or store telemetry itself. You need separate infrastructure for the data layer: Prometheus for metrics, Loki for logs, Tempo for traces. A complete observability stack requires provisioning and maintaining these components alongside Grafana. For teams starting from scratch, the full setup involves moderate complexity.
- -3,300+ open issues reflect a broad surface areaWith over 3,300 open GitHub issues, the tracker reflects the breadth of integrations and configurations the community uses. Feature requests, edge-case bugs, and plugin compatibility issues accumulate. Most day-to-day observability use cases work well, but niche data source integrations or advanced alerting scenarios may require workarounds.
Grafana vs alternatives#
Grafana vs Datadog
Grafana and Datadog overlap on dashboards and alerting, but they are built on opposite deployment models. Grafana is the visualization layer you run yourself, connected to data sources you already operate. Datadog is a fully managed monitoring platform that collects, stores, and visualizes telemetry from agents installed on your hosts.
| Feature | Grafana | Datadog |
|---|---|---|
| License | AGPL-3.0 | Proprietary |
| Self-hosting | Yes | No |
| Data storage | Your own backends | Datadog cloud |
| Pricing | Free self-hosted | Per-host, usage-based |
| Built-in APM | No (needs Tempo or Jaeger) | Yes |
Grafana is the stronger choice when you have existing data infrastructure (Prometheus, Loki, Elasticsearch) and want dashboards without sending all your telemetry to a third-party cloud. Datadog is worth its cost when you need out-of-the-box APM with automatic code tracing, built-in anomaly detection with no infrastructure to manage, or a single vendor contract covering security, logging, and monitoring.
Grafana vs New Relic
New Relic targets full-stack observability as a managed service, including application performance monitoring, browser monitoring, and infrastructure monitoring in one platform. Grafana requires you to bring your own data backends for each signal type.
| Feature | Grafana | New Relic |
|---|---|---|
| License | AGPL-3.0 | Proprietary |
| Self-hosting | Yes | No |
| Free tier | Full open source | 100GB/month data ingest |
| APM | Via OpenTelemetry backend | Native, with language agents |
Teams already invested in the Prometheus and OpenTelemetry ecosystem will find Grafana the more natural fit: their existing instrumentation works without rerouting to New Relic's ingest. New Relic's free tier is genuinely useful for smaller teams that want a managed platform with no infrastructure overhead, and its automated correlation features are ahead of what the open source Grafana stack provides out of the box.
Grafana vs Dynatrace
Dynatrace positions itself as an AI-powered observability platform with automatic dependency mapping and root-cause analysis. It installs a single agent per host that discovers topology automatically. Grafana requires you to configure what to collect and how to connect data sources, which gives more control but demands more setup.
Grafana is preferable when your team wants full control over the observability stack, prefers query-driven investigation, and runs Kubernetes or cloud infrastructure with existing Prometheus exporters. Dynatrace is better suited for large enterprise environments where reducing manual instrumentation effort justifies the license cost, or where automatic infrastructure topology mapping is a hard requirement.
Install and self-host#
Self-hosting Grafana uses Docker with a single container and no external database required for a basic setup.
```bash
docker run -d -p 3000:3000 grafana/grafana
```What it's built on#
- Languages
- GoJavaScriptTypeScript
- Frameworks
- React
- Databases
- MySQLPostgreSQL
- Search
- Elasticsearch
- Tooling
- esbuildRollupWebpack
FAQ#
Is Grafana free to use?
The open source core of Grafana is AGPL-3.0 licensed and free to self-host. You can run it on your own server at no cost. Grafana Labs also offers a managed cloud tier (Grafana Cloud) with a free-forever plan for small deployments and paid plans for larger volumes of metrics, logs, and traces.
What data sources does Grafana support?
Grafana includes official plugins for over 100 data sources, including Prometheus, Loki, Elasticsearch, InfluxDB, Tempo, MySQL, Postgres, CloudWatch, Azure Monitor, Google Cloud Monitoring, and more. The plugin architecture allows community-built connectors for additional systems. Any system with a queryable API can be connected via a custom data source plugin.
How do I install Grafana?
The fastest path is Docker: running 'docker run -d -p 3000:3000 grafana/grafana' starts a Grafana instance accessible on port 3000. Official packages are also available for Ubuntu, Debian, RHEL, and Windows. For Kubernetes deployments, Grafana Labs maintains an official Helm chart. Full installation documentation is at grafana.com/docs.
What is the difference between Grafana and Grafana Cloud?
Grafana (the open source project) is the visualization and alerting layer you run on your own infrastructure. Grafana Cloud is the managed service offered by Grafana Labs: they handle hosting, scaling, and data storage for metrics, logs, and traces. The core dashboarding experience is the same. The difference is who manages the backend data infrastructure and who pays for the servers.
Does Grafana work with Prometheus?
Yes. Prometheus is one of Grafana's native and most widely used data sources. Grafana includes a built-in Prometheus data source plugin that supports PromQL queries, label filtering, and alerting rules. Most Grafana users pair it with Prometheus for metrics collection and Loki or Elasticsearch for log aggregation.
Similar open-source tools#
Prometheus
ClawTrace
Visualize agent execution trees and track token costs per step
Stakpak
Open source agent that automates app deployment and monitoring
OpenObserve
Self-hosted observability for logs, metrics, and traces
Coroot
Instant observability with no-code setup.
Netdata
Real-time infrastructure monitoring with zero configuration