Grafana

What Is Grafana?

When the platform engineering team at a mid-sized European fintech hit 200 microservices last year, their monitoring setup was barely holding together. Datadog bills had crossed $15,000 per month. Half the team used CloudWatch, the other half had Prometheus instances nobody maintained. Alerting was a chaos of Slack channels and PagerDuty noise. They needed a single pane of glass — one that could unify their metrics, logs, and traces without locking them into another expensive proprietary vendor.

They chose Grafana. Within three weeks, the team had a unified observability platform pulling from Prometheus, Loki, and Tempo, with dashboards that gave every engineer — from backend to SRE — a real-time view of the entire system. Their monitoring costs dropped by roughly 60%.

This story is not unusual. Grafana, originally created in 2014 by Swedish developer Torkel Odegaard as a fork of Kibana, has become the de facto standard for observability dashboards. The open-source project now has over 60,000 GitHub stars, supports 150+ data source plugins, and is deployed at organisations ranging from two-person startups to some of the largest enterprises on the planet.

Grafana Labs, the company behind the project, is headquartered in New York (legally Raintank Inc.) with approximately 1,800 employees across 20+ countries. As of early 2026, the company is reportedly raising funding at a $9 billion valuation, having previously raised over $800 million. While the founder is Swedish and the project has deep European roots, the corporate entity is American — a fact that matters for EU compliance considerations.

Key Features

Dashboard Builder

Grafana's core strength is its dashboard builder. With 30+ visualisation panel types — time series, bar gauges, heatmaps, histograms, geomap, node graphs, flame graphs, and more — you can represent virtually any metric visually. Template variables let you create dynamic, reusable dashboards where a single panel definition can render differently based on environment, service, or region. Dashboard linking lets you drill from a high-level overview down to granular detail in a few clicks.

The dashboard JSON model is fully exportable and versionable. Teams routinely store dashboard definitions in Git, provision them via Terraform, and deploy them as code. Grafana's "observability as code" approach means your monitoring infrastructure is as reproducible as your application infrastructure.

Unified Alerting

Grafana's alerting system lets you define alert rules across any connected data source — Prometheus metrics, Loki log patterns, or Tempo trace conditions. Alert rules support multi-dimensional evaluation (one rule can fire separately for each service), notification policies route alerts to the right team, and silences suppress known issues during maintenance windows.

Grafana OnCall extends this with on-call scheduling, escalation chains, and mobile push notifications. The combination of alerting and OnCall gives teams an integrated incident management workflow without needing a separate tool like PagerDuty — though Grafana integrates with PagerDuty too, for teams that prefer it.

The LGTM Stack

Grafana Labs maintains four open-source backends that together form the LGTM stack: Loki for logs, Grafana for dashboards, Tempo for traces, and Mimir for metrics. Each is designed to be horizontally scalable, cost-efficient, and compatible with open standards like OpenTelemetry.

Loki, in particular, is a standout. Unlike Elasticsearch, it indexes only metadata (labels) rather than full log content, which makes it dramatically cheaper to operate at scale. For teams drowning in log storage costs, switching from ELK to Loki can reduce expenses by an order of magnitude.

150+ Data Source Plugins

Grafana connects to virtually anything that emits data. Native plugins cover Prometheus, Elasticsearch, OpenSearch, InfluxDB, PostgreSQL, MySQL, CloudWatch, Azure Monitor, Google Cloud Monitoring, and dozens more. The plugin catalogue includes community-contributed connectors for niche databases, IoT platforms, and business intelligence tools. If your data exists somewhere, Grafana can probably visualise it.

Grafana Cloud

For teams that do not want to self-host, Grafana Cloud provides a fully managed observability platform. It bundles managed Prometheus (Mimir), managed Loki, managed Tempo, and the Grafana dashboard — along with newer services like Application Observability (automatic distributed tracing), Grafana IRM (incident response management), and k6 Cloud for load testing.

The free tier is genuinely generous: 10,000 metric series, 50GB of logs, 50GB of traces, and 3 users with 14-day retention. Many small teams run their entire observability stack on the free tier indefinitely.

AI Assistant

Grafana Assistant, introduced in late 2025, is an LLM-powered agent built directly into Grafana Cloud. It can help troubleshoot dashboard issues, generate PromQL queries from natural language descriptions, explain alert conditions, and suggest optimisations. It is a useful acceleration tool for teams ramping up on Grafana's query languages, though experienced users will still prefer writing queries directly.

Pricing

Grafana's pricing model is split between the open-source self-hosted option (free forever) and Grafana Cloud's tiered plans.

Free includes 10K metrics series, 50GB logs, 50GB traces, 3 users, and 14-day retention. For small teams and side projects, this is often enough.

Cloud Pro starts at $19/month as a platform fee, plus usage-based charges: $6.50 per 1,000 additional metric series, $0.55 per GB of additional logs or traces, and $8 per additional active user. Grafana uses 95th percentile billing for metrics, which means temporary spikes do not inflate your bill. This is a meaningful advantage over Datadog's per-host pricing model, which penalises auto-scaling.

Enterprise requires a minimum $25,000 annual commitment and adds dedicated support, SLAs, custom retention, and enterprise plugins for connectors like ServiceNow, Splunk, and Oracle.

The self-hosted option deserves emphasis. Running Grafana, Prometheus, Loki, and Tempo on your own infrastructure costs only the compute and storage you provision. For a mid-sized deployment on European cloud providers like Hetzner or OVHcloud, this can be as little as EUR 50-200/month — a fraction of equivalent Datadog or New Relic costs. The trade-off is operational complexity: you are responsible for upgrades, scaling, and reliability.

EU Compliance & Privacy

This is where honesty matters. Grafana was created in Sweden by a Swedish developer, and the project has strong European heritage. But Grafana Labs, the company, is legally Raintank Inc., headquartered at 165 Broadway, New York. It is a US company subject to US jurisdiction, including potential FISA Section 702 and CLOUD Act data requests.

For Grafana Cloud, the company participates in the EU-US Data Privacy Framework and offers an EU region hosted on GCP Belgium. It holds SOC 2 Type II and ISO 27001 certifications, offers a GDPR Data Processor Agreement, and is PCI DSS compliant. These are solid certifications — but they do not change the fundamental jurisdictional reality. Data stored on Grafana Cloud, even in the EU region, is controlled by a US legal entity.

For organisations with strict data sovereignty requirements — government agencies, healthcare providers, financial regulators — the self-hosted path is the answer. Running Grafana OSS on European infrastructure (Hetzner, OVHcloud, Exoscale, Scaleway) keeps all observability data under EU jurisdiction with zero US involvement. This is one of the genuine advantages of Grafana's open-source model: you are not dependent on Grafana Labs for compliance.

We score EU compliance at 5.5/10 for Grafana Cloud specifically. The US parent company is a real limitation for regulated European organisations. Self-hosted Grafana would score significantly higher, but the rating reflects the product as commercially offered.

Who It's Best For

Platform and SRE teams building observability stacks across heterogeneous infrastructure. Grafana's data source agnosticism means you can unify AWS, Azure, on-premise, and Kubernetes monitoring into one interface.

Cost-conscious engineering organisations looking to escape Datadog or New Relic vendor lock-in. Self-hosted Grafana with the LGTM stack can reduce observability costs by 50-80% compared to proprietary alternatives.

European companies with data sovereignty requirements who can self-host on EU infrastructure. The open-source model provides full compliance control that no US-hosted SaaS can match.

DevOps teams standardising on OpenTelemetry who need a visualisation and alerting layer that natively supports OTel signals across metrics, logs, and traces.

The Verdict

Grafana is the most capable open-source observability platform available. Its dashboard builder is unmatched, its data source ecosystem is enormous, and its cost model — whether self-hosted or on the generous Cloud free tier — makes enterprise-grade monitoring accessible to teams of any size. The EU compliance picture is mixed: the US parent company is a genuine concern for regulated industries, but the open-source self-hosting option provides an escape hatch that proprietary competitors simply cannot offer. For European teams willing to invest in operational capability, Grafana delivers world-class observability without vendor lock-in.

Frequently Asked Questions

Is Grafana free to use?

Yes, in two ways. Grafana OSS is open-source under AGPLv3 and free to self-host with no usage limits. Grafana Cloud also offers a free tier with 10K metrics series, 50GB logs, 50GB traces, and 3 users — no credit card required.

Can I run Grafana entirely within the EU?

Yes. Self-hosting Grafana on European cloud providers (Hetzner, OVHcloud, Scaleway, Exoscale) keeps all data under EU jurisdiction. For Grafana Cloud, an EU region is available on GCP Belgium, though the controlling entity remains a US company.

How does Grafana compare to Datadog?

Grafana is open-source and vendor-agnostic, connecting to 150+ data sources. Datadog is proprietary and all-in-one. Grafana is dramatically cheaper at scale, especially self-hosted, but requires more setup and operational knowledge. Datadog provides a more polished out-of-the-box experience with stronger APM and log management workflows.

What is the LGTM stack?

LGTM stands for Loki (logs), Grafana (dashboards), Tempo (traces), and Mimir (metrics). These four open-source projects from Grafana Labs form a complete, horizontally scalable observability backend that is compatible with OpenTelemetry and Prometheus standards.

Do I need to know PromQL to use Grafana?

Not necessarily, but it helps significantly. Grafana's query editor provides auto-completion and visual builders for simpler queries. The new Grafana Assistant can generate queries from natural language. However, for complex dashboards and alerting rules, learning PromQL (for metrics) and LogQL (for logs) will unlock Grafana's full power.