The evolution of the Grafana observability stack

Grafana Labs began in 2014 as a provider of open-source visualization. The company reached a $6 billion valuation in August 2024 following a $270 million Series D extension led by Lightspeed Venture Partners. This funding followed a $240 million Series D round and a $220 million Series C raise that valued the company at $3 billion. The software supports more than 10 million users and over 900,000 active installations worldwide. While the company started as a dashboarding tool, it expanded into a full observability suite including Prometheus for metrics, Loki for logs, and Tempo for traces. The company also developed Mimir, a time-series database that is 40 times faster than Cortex. Last year, the company acquired Asserts.ai to improve its application observability capabilities. Companies such as 7AI, TeamSystem, and Zama now use Grafana Cloud to manage their AI-era infrastructure. Zama specifically uses Grafana Assistant to help engineers investigate issues and accelerate debugging. The company has grown its paying customer base to more than 5,000 companies, including Dell Technologies, Microsoft, and Citigroup. Revenue for the company has risen to more than $250 million. The Grafana Enterprise Stack includes Enterprise Metrics for delivering Prometheus-as-a-service and Enterprise Logs for scalable log aggregation.

Architecture and data control

Engineering leaders choosing between Datadog, New Relic, and Grafana must decide if they want to own their stack or rent a service. Datadog and New Relic function as SaaS-based suites that use proprietary agents to scrape data from infrastructure. These tools reduce the mean time to resolution by normalizing data formats automatically, whereas Grafana requires infrastructure teams to manage storage backends, retention policies, and the query performance of underlying data sources. If your organization operates in healthcare or finance, the Grafana stack provides a way to keep telemetry data within a private VPC to meet strict data residency requirements. New Relic provides deeper application performance monitoring because its agents instrument code at the runtime level to show specific function call timings. You might find the manual configuration of OpenTelemetry in Grafana more tedious than the out-of-the-box experience in New Relic. For high-performance front-end applications, Datadog maintains a single session for up to four hours, while a New Relic session ends if the application remains unused for 15 minutes. Datadog provides a security solution through its Cloud SIEM, while New Relic does not provide a security monitoring feature yet. While New Relic uses NRQL to query its NRDB database, Grafana users rely on PromQL for metrics and LogQL for logs. Datadog uses honeycomb charts to display infrastructure metrics. New Relic focuses on showing all traces in groups to help filter root spans. You should also factor the cost of engineering time into your total cost of ownership. If a developer spends 10 hours a month maintaining a self-hosted Grafana stack, that labor cost can exceed a premium managed service fee.

Pricing and the decision for engineering leaders

Pricing models drive most procurement decisions for modern engineering teams. New Relic uses a model based on data ingestion and user seats, which can cause bills to spike when a deployment increases telemetry volume. New Relic provides a 100GB free tier, but costs increase once teams exceed these limits, including a $49 monthly fee for each full-platform user. Datadog uses per-host pricing starting at $15 per host for infrastructure monitoring and $31 per host for APM. Grafana Cloud provides a middle path with a usage-based model that eliminates the per-user fees that make New Relic expensive for growing teams. Grafana Cloud’s free tier includes 10,000 metrics series, 50GB of logs, and 50GB of traces. For many small to mid-sized teams, Grafana Cloud’s pricing is lower than New Relic’s because Loki’s label-only indexing keeps storage costs low. New Relic’s 100GB free tier is the best in the industry for a single user, but a five-person team will face significant costs as they add seats.

I recommend Grafana Cloud for teams that need to control costs and avoid vendor lock-in through OpenTelemetry. Does the operational burden of managing Prometheus and Loki outweigh the cost savings for your specific team?