Learn how your cost is calculated
Grafana Cloud uses different billing methodologies depending on the product to offer predictable pricing. Understanding these methods helps you read the Usage tab and decide where to optimize.
95th percentile billing for metrics
Grafana Cloud meters metrics on two components:
- Active series: Unique time series that have received a data point within the last 20 minutes.
- Data points per minute (DPM): How frequently data is sent to Grafana Cloud.
Your billable metrics usage is the maximum of either active series or total DPM normalized by the included DPM per series:
usage = max(active_series, total_dpm / included_dpm)
Grafana Cloud Pro includes 1 DPM per active series. If your average DPM per active series exceeds that, billing is based on total DPM.
The 95th percentile advantage
Grafana Cloud bills metrics on the 95th percentile of usage across the billing period, which:
- Forgives temporary spikes during incidents or deployments.
- Provides predictable monthly costs based on typical usage.
- Ignores the top 5% of usage time each month (approximately 36 hours).
So if you normally run 6,000 active series but spike to 30,000 for 24 hours in a month, you’re still billed near the 6,000 series rate.
Scrape interval impact on costs
Your scrape interval directly affects DPM and cost:
- 15-second interval: 4 data points per minute (higher cost)
- 30-second interval: 2 data points per minute (medium cost)
- 60-second interval: 1 data point per minute (lower cost)
A common approach is to use a 1-minute scrape interval as the global default and only increase frequency for jobs that need higher resolution. Refer to Reduce metrics costs by adjusting your data points per minute (DPM).
Volume-based billing for logs, traces, and profiles
Grafana Cloud Logs, Traces, and Profiles share a usage model based on the volume of data, measured across three units:
- Process: Data received and optimized as it arrives.
- Write: Data ingested and stored.
- Retain: Data kept beyond the included retention window.
For the full breakdown, refer to Understand your Grafana Cloud Logs, Traces, and Profiles invoices.
Specialized product billing
Other products have tailored billing models:
- Application Observability: Billed by host hours. Refer to Understand your Application Observability invoice.
- Frontend Observability: Billed by ingested sessions. Refer to Understand your Frontend Observability invoice.
- Incident Response & Management (IRM): Billed by monthly active users. Refer to Understand your Grafana Cloud IRM invoice.
- Kubernetes Monitoring: Billed by host hours and container hours. Refer to Understand your Kubernetes Monitoring invoice.
- Synthetic Monitoring: Billed by test executions. Refer to Understand your Synthetic Monitoring invoice.
- Performance Testing (k6): Billed by Virtual User Hours (VUh). Refer to Understand your Performance Testing invoice.
Optimization strategies
Understanding billing helps you optimize cost:
- Apply Adaptive Telemetry recommendations to drop low-value data.
- Use the Cardinality tab to find high-cost metrics.
- Adjust scrape intervals based on actual monitoring needs.
- Review unused metrics for potential removal.
For more, refer to Adaptive Telemetry and Analyze and reduce costs.
In the next milestone, you’ll create a usage alert to proactively monitor your Grafana Cloud costs.
More to explore (optional)
At this point in your journey, you can explore the following topics: