Application Performance Monitoring
Fast & Cost-Effective APM with AI-based sampling. Runs On-Prem with no traces data sent out of your cloud.
Easy log management with fast search & data privacy at the core. No log data leaves your cloud. Ever.
Track hosts, databases, containers, and cloud services in real time — with unlimited retention and full data control.
Track every user journey in real time. Get instant visibility into performance bottlenecks with 4X faster, on-prem RUM.
Prevent downtime before it happens with proactive, scripted tests. Simulate user journeys and detect issues before customers are impacted.
Catch every error in real time. Aggregate, analyze, and resolve issues 4× faster with CubeAPM.
Track Kubernetes clusters, nodes, and pods with real-time performance insights
Application Performance Monitoring
Fast & Cost-Effective APM with AI-based sampling. Runs On-Prem with no traces data sent out of your cloud.
Easy log management with fast search & data privacy at the core. No log data leaves your cloud. Ever.
Track hosts, databases, containers, and cloud services in real time — with unlimited retention and full data control.
Track every user journey in real time. Get instant visibility into performance bottlenecks with 4X faster, on-prem RUM.
Prevent downtime before it happens with proactive, scripted tests. Simulate user journeys and detect issues before customers are impacted.
Catch every error in real time. Aggregate, analyze, and resolve issues 4× faster with CubeAPM.
Track Kubernetes clusters, nodes, and pods with real-time performance insights
Grafana Pricing Calculator
Grafana Cloud Pricing Calculator
Enter your usage to get an instant monthly estimate - and see how CubeAPM's on-prem pricing compares.
Typical ranges: Small cluster 20k-100k · Mid-size K8s 200k-1M · Enterprise 5M+
~15-40 series/container is common for a well-instrumented Kubernetes cluster
scrape_interval config.
CubeAPM is built for teams with $1,000+/month in observability spend.
See where CubeAPM wins (Mid-Market) →
This estimate uses conservative assumptions.
Grafana Cloud cost breakdown
Unlimited users · Unlimited retention
| Component | Grafana Cloud | CubeAPM |
|---|
What's included with CubeAPM
| Grafana Cloud | CubeAPM | |
|---|---|---|
| Billing dimensions | 8+ independent meters | $0.15/GB |
| Metrics pricing | Per series (cardinality risk) | Included in $0.15/GB rate |
| Data retention | 14 days free, metered paid | Unlimited included |
| Users | Per-seat above 3 | Unlimited |
| Cardinality risk | Yes - label explosion = bill explosion | No |
| Infrastructure | Grafana-managed SaaS | Runs in your own cloud |
Metrics: 95th-percentile billing assumed. "Active series" = unique metric_name + label set. Low-resolution rate used ($6.50/1k); high-resolution is $16/1k.
Logs / Traces / Profiles: $0.05/GB process + $0.40/GB write = $0.45/GB above 50 GB free per signal. 30-day Pro retention included - no extra retain charge assumed. 100% write-through assumed (no Adaptive Logs filtering configured).
Kubernetes Monitoring: 744 hours/month (31-day month). 1 host = 1 Kubernetes node, not a pod.
CubeAPM: $0.15/GB platform fee. ~$0.02/GB estimated cloud infra overhead. $1,000/month minimum. GB equivalents are estimates - actual volume varies by instrumentation depth.
Not affiliated with or endorsed by Grafana Labs, Inc.
Why use this Grafana Cloud pricing calculator?
Grafana Cloud pricing is usage-based and spans multiple independent dimensions: metrics series, log and trace ingestion volume, host hours for Kubernetes and Application Observability, database host hours, per-user charges for dashboards and incident management, and test executions for synthetics. Each dimension is billed separately, and changes in any one of them can materially shift your monthly bill without being obvious upfront.
This calculator helps you:
- Estimate your Grafana Cloud spend before committing to a plan or scaling workloads
- Identify which dimensions (metrics cardinality, log volume, host hours, or active users) drive the most cost in your environment
- Model cost scenarios across different team sizes and observability coverage levels
- Compare Grafana Cloud’s usage-based pricing against a more predictable alternative using the same assumptions
The goal is not to predict an exact bill. Actual charges depend on usage patterns, retention choices, and any volume discounts. But it gives you a realistic cost range based on how Grafana Cloud pricing works in practice.
Why Grafana Cloud's pricing becomes complex at scale
Grafana Cloud is composable by design: you enable only the products you need and pay for what you use. In practice, as environments grow, costs accumulate across a large number of independent billing units, and the interactions between them are not always intuitive.
Plan tiers
Grafana Cloud has three plan tiers:
- Free: $0, always. Includes limited usage across all products with 14-day retention and community support.
- Pro: From $19/month + usage. Longer retention, 8×5 email support, and pay-as-you-go above the free tier included usage.
- Enterprise: Starts at $25,000/year spend commit. Custom retention, premium support, and deployment flexibility.
The $19/month Pro fee is a platform fee. The actual bill grows on top of it as usage scales across each product.
Metrics
Grafana Cloud charges for metrics based on billable series, not the count of metric names. A single metric with high-cardinality labels generates one series per unique label combination. For example, a metric labelled by Kubernetes pod name across 500 pods creates 500 billable series from a single metric name.
Pricing:
- Free tier: 10,000 active series/month, 14-day retention
- Pro: $6.50 per 1,000 series/month (13-month retention included)
- Enterprise: from $3 per 1,000 series on annual commit
This catches teams off guard because Prometheus scrapers with many targets, Kubernetes environments with per-pod labels, and new exporters can silently add thousands of series as workloads scale.
Logs
Grafana Cloud logs are powered by Grafana Loki. Billing is split across three operations per GB ingested:
- Process: $0.05/GB (receiving and optimizing the incoming signal)
- Write: $0.40/GB (writing useful data to storage)
- Retain: $0.10/GB (keeping stored data over time)
Free tier includes 50 GB ingested/month with 14-day retention. Pro includes 50 GB/month with 30-day retention, then pay as you go. Extending retention beyond 30 days requires contacting Grafana for custom pricing.
The three-part billing structure means even a simple decision like increasing log verbosity shows up as three separate line items on your invoice.
Traces
Grafana Cloud traces (powered by Grafana Tempo) use the same three-part billing structure as logs:
- Process: $0.05/GB
- Write: $0.40/GB
- Retain: $0.10/GB
Free tier: 50 GB/month, 14-day retention. Pro: 50 GB/month included, 30-day retention, then pay as you go. Enterprise: custom pricing on annual commit.
Without sampling configured, a service handling high request volume can exceed the 50 GB free tier quickly. Tail-based sampling (keeping traces for errors and high-latency requests only) is the primary lever for controlling trace costs.
Profiles
Continuous profiling data (powered by Grafana Pyroscope) follows the same pricing as logs and traces:
- Process: $0.05/GB
- Write: $0.40/GB
- Retain: $0.10/GB
Free tier: 50 GB/month, 14-day retention. Pro: 50 GB/month included, 30-day retention, then pay as you go. Enterprise: custom pricing on annual commit.
Kubernetes Monitoring
Kubernetes Monitoring is billed by host hours and container hours, not by telemetry volume alone. A host is any physical or virtual OS instance sending observability signals; it is considered active if it has sent a signal in the last 15 minutes.
Pricing:
- Free tier: 2,232 host hours + 37,944 container hours/month
- Pro: $0.01/host hour (~$7.20/host/month) + $0.0007/container hour (~$0.50/container/month)
- Enterprise: custom pricing on annual commit
The $19/month platform fee includes the free tier allotments, then charges apply on top. Telemetry generated by monitored hosts (metrics, logs, traces) is billed separately at standard telemetry rates.
This creates a compound cost structure: each Kubernetes node contributes to host-hour charges AND to telemetry volume charges simultaneously.
Application Observability
Application Observability (APM) is billed per host hour. A host is active if it has sent a signal in the last 15 minutes.
Pricing:
- Free tier: 2,232 host hours/month (~3 hosts running all month)
- Pro: $0.025/host hour (~$18/host/month), including telemetry generated
- Enterprise: custom pricing on annual commit
As with Kubernetes Monitoring, host-hour charges for Application Observability stack on top of standard telemetry charges (metrics, logs, traces) generated by those same hosts.
Frontend Observability
Frontend Observability provides real user monitoring (RUM) and is billed per session. A session is the time a user spends in the application; it ends after 4 hours maximum or 15 minutes of inactivity.
Pricing:
- Free tier: 50,000 sessions/month
- Pro: $0.75 per 1,000 sessions, pay as you go above the free tier
- Enterprise: custom pricing on annual commit
Session counts scale directly with user traffic, so high-traffic web applications can see Frontend Observability costs grow quickly during peak periods.
Database Observability
Database Observability is billed per database host hour. Each database server instance counts as one database host, including individual nodes in a cluster.
Pricing:
- Free tier: 2,232 database host hours/month (~3 database hosts running all month). Telemetry counts toward standard free tier allowances.
- Pro: $0.07/host hour (~$51/database host/month). Telemetry billed separately at standard rates.
- Enterprise: custom pricing on annual commit
Database Observability charges are additive to the telemetry (metrics and logs) those database hosts generate. Running monitoring on 5 database hosts costs roughly $255/month in host-hour charges alone, before accounting for the metrics and logs they produce.
Synthetics
Synthetic monitoring is billed per test execution. One test execution equals a synthetic test running in one probe location for one minute of run time.
Pricing:
- Free tier: 100,000 API test executions + 10,000 browser test executions/month
- Pro API tests: $5 per 10,000 test executions
- Pro browser tests: $50 per 10,000 test executions
- Enterprise: custom pricing on annual commit
To estimate executions: multiply number of probe locations x number of tests x test duration (minutes) x (43,200 / test frequency in minutes). A single test running every minute from 5 locations generates approximately 216,000 executions/month, well above the free tier limit.
Performance Testing (k6)
Load and performance testing is billed in virtual user hours (VUh). VUh is calculated as: (maximum VUs x test duration in minutes) / 60.
Pricing:
- Free tier: 500 VUh/month
- Pro: starts at $0.15/VUh, with volume discounts at higher usage
- Enterprise: as low as $0.05/VUh on annual commit
Grafana Visualization
Grafana dashboards are billed per active user. Any user who logs in during the billing month counts as active.
Pricing:
- Free tier: 3 active users/month
- Pro: $8/active user/month (without Enterprise plugins); $55/active user/month (with Enterprise plugins)
- Enterprise: custom pricing on annual commit
Incident Response and Management (IRM)
IRM (on-call scheduling, alerting, and incident management) is billed per active IRM user. A user is active if they are included in on-call schedules, receive a page, or create/update an incident during the billing month.
Pricing:
- Free tier: 3 active IRM users/month
- Pro: $20/active IRM user/month
- Enterprise: custom pricing on annual commit
Grafana Assistant (AI)
The AI copilot is billed per active AI user. A user is active if they send a message or use an Assistant feature during the billing month.
Pricing:
- Free tier: 3 active AI users/month (200 messages/user/month)
- Pro: $20/active AI user/month (2,000 messages/user/month included)
- Enterprise: custom pricing on annual commit
What typically causes Grafana Cloud bills to spike
Most unexpected cost increases come from a recognizable set of patterns:
- High-cardinality labels on metrics: Adding a label like user_id, pod, or request_id multiplies billable series by the cardinality of that label across all active instances
- Verbose or unfiltered log shipping: Sending debug or info logs from all services in production without filtering or sampling
- Kubernetes scale-out events: Temporary autoscaling spikes increase both host-hour charges and telemetry volume simultaneously
- Enabling new products without cost modeling: Turning on Application Observability, Database Observability, or IRM adds new per-host or per-user charges that stack on top of existing telemetry spend
- No trace sampling configured: Deploying OpenTelemetry across services without head-based or tail-based sampling floods trace ingestion rapidly
- Growing team access to dashboards and IRM: Per-user charges for Visualization and IRM grow quietly as more engineers are granted access
How to manage Grafana Cloud costs
Reduce metrics series count with Adaptive Metrics
Grafana Cloud’s Adaptive Metrics feature identifies unused or over-specified time series and recommends aggregating or dropping them. For teams with large Prometheus deployments, this is typically the highest-leverage cost control available. Grafana Labs publishes case studies showing up to 80% reduction in billable series through this feature.
Filter logs at the source with Adaptive Logs
Adaptive Logs identifies frequently ingested log patterns that are rarely queried and recommends filters to drop them before ingestion. Reducing log volume by 30-50% has a direct, proportional impact on the Process, Write, and Retain charges.
Configure trace sampling
Trace volume is directly controlled by sampling at the SDK or collector level. Tail-based sampling (keeping traces for errors and high-latency requests only, rather than randomly) preserves signal while reducing ingested volume more efficiently than head-based sampling.
Scope host-hour monitoring to production
Kubernetes Monitoring, Application Observability, and Database Observability all bill by active host hours. Limiting monitoring to production namespaces rather than all environments reduces host-hour charges without reducing coverage where it matters most.
Audit active users regularly
Per-user charges for Visualization, IRM, and the AI Assistant grow quietly as more team members are granted access. Periodic audits to deactivate users who no longer need access can reduce user-based costs meaningfully in larger organizations.
Use the free tier for non-production workloads
The free tier is functional: 10,000 active series, 50 GB each of logs, traces, and profiles, 3 active users, 14-day retention. Routing dev and staging traffic to a separate free-tier workspace keeps those costs entirely out of your primary bill.
When cost optimization becomes ongoing work
The measures above are effective but require continuous attention as systems evolve. High-cardinality labels reappear when new engineers instrument services without billing awareness. Log verbosity creeps back up after debugging sessions. New Grafana Cloud products get enabled by individual teams without FinOps coordination.
At a larger scale, managing observability spend inside a multi-dimensional usage-based model becomes a recurring operational task. This is often when teams begin evaluating whether structural cost controls enforced by architecture better fit their needs than process-based optimization.
Grafana Cloud Pricing vs CubeAPM: A Structural Difference
The fundamental difference between Grafana Cloud and CubeAPM is not features. It is where the platform runs and how costs are bounded.
Grafana Cloud is a fully managed SaaS platform. All telemetry (metrics, logs, traces, profiles) is shipped to Grafana Labs’ infrastructure and billed based on volume sent, hosts monitored, and users active. Costs grow with telemetry, and telemetry grows with systems.
CubeAPM runs inside your own cloud account (AWS, GCP, Azure) or on-premises infrastructure. Telemetry data never leaves your environment. You pay for the compute and storage you provision in your own cloud, not for usage metered by a SaaS vendor. Cost growth is tied to infrastructure planning, not instrumentation depth.
| Area | Grafana Cloud | CubeAPM |
|---|---|---|
| Deployment | Fully managed SaaS | Self-hosted in your own cloud |
| Pricing model | Per series + per GB (Process/Write/Retain) + per host hour + per user | Per GB ingested (logs, traces, metrics) |
| Cost predictability | Varies across multiple independent dimensions | Stable; tied to infrastructure provisioning |
| Data residency | Grafana Labs infrastructure | Your cloud account; data never leaves |
| Metrics billing | Billable series (cardinality-sensitive) | GB ingested |
| Log and trace retention | 30 days default; custom via Enterprise contract | No retention fees; storage is your own infrastructure cost |
| Per-user charges | $8/user (Visualization), $20/user (IRM), $20/user (AI Assistant) | Unlimited users included |
| Data egress | Telemetry exits your VPC to Grafana; cloud egress applies | Zero egress; CubeAPM runs in-VPC |
| Support | Community (Free), 8x5 email (Pro), custom (Enterprise) | Included with plan |
With CubeAPM, the cost drivers are the infrastructure you provision, your retention choices, and your sampling strategy. There are no per-series charges, no per-user fees, no separate APM licensing, and no data egress costs. Instrumentation decisions are engineering decisions, not billing events.
Grafana Cloud remains a strong choice for teams that want a fully managed experience with no operational overhead, and that are comfortable with usage-based pricing in exchange for not managing infrastructure. The tradeoff is real: CubeAPM requires provisioning and operating infrastructure inside your own account, which has its own cost in engineering time.
The calculator above is designed to help you see both sides of that tradeoff clearly, using your own usage estimates.
Disclaimer: Pricing based on Grafana Cloud public pricing at grafana.com/pricing, verified May 2026. Actual costs vary based on usage patterns, volume discounts, and enterprise contract terms. This calculator is not affiliated with or endorsed by Grafana Labs.