Observability Tools Pricing Comparison [2026]

Vladimir Mihailenco
February 07, 2026
12 min read

Observability costs can quietly become one of the largest line items in your infrastructure budget. Trace volumes grow with traffic, log ingestion scales with services, and metrics multiply with every new host and deployment. Without a clear understanding of how each vendor charges, teams often face surprise bills that dwarf initial estimates.

This guide compares monthly costs for five popular observability platforms — Datadog, New Relic, Grafana Cloud, SigNoz, and Uptrace — across three realistic team sizes. All estimates use published list pricing with annual commitments where available, and we show the math so you can adapt the numbers to your own usage.

If you are evaluating top APM tools, pricing should be a key part of the decision alongside features, ease of setup, and long-term flexibility.

Quick Comparison

Team SizeUptraceSigNozGrafana CloudNew RelicDatadog
Large (150 eng)$2,426$12,240$20,955$24,635$42,468
Midsize (65 eng)$1,349$5,580$9,525$11,477$20,988
Small (25 eng)$344$1,374$2,318$3,497$4,856

All figures are estimated monthly costs in USD based on published unit pricing. Uptrace totals include estimated AWS/GCP egress fees for sending data to the Hetzner datacenter (see breakdown). Actual costs vary by usage patterns, retention settings, and negotiated discounts.

Team Size Definitions

Each tier represents a realistic engineering organization with full-stack observability (traces, logs, and metrics):

LargeMidsizeSmall
Engineers1506525
APM Hosts22512520
Trace Volume10 TB/month5 TB/month1 TB/month
Log Volume20 TB/month10 TB/month2.5 TB/month
Infrastructure Hosts35020050
Active Timeseries750,000250,00075,000

These profiles reflect common ratios: roughly 1.5 APM hosts per engineer, 2–3 infrastructure hosts per engineer, and log volumes that typically exceed trace volumes by 2x.

How Each Tool Charges

Datadog

Datadog uses a multi-dimensional pricing model. You pay separately for host licenses, data ingestion, event indexing, and custom metrics. Annual pricing (per host/month):

  • Infrastructure Enterprise: $23/host
  • APM Enterprise: $40/host
  • Log Ingestion: $0.10/GB
  • Log Indexing (15-day retention): $1.70/million events
  • Ingested Spans: $0.10/GB beyond 150 GB/host allotment
  • Indexed Spans (15-day retention): $1.70/million spans beyond 1M/host allotment
  • Custom Metrics: $0.01/metric/month beyond 200 per Enterprise host

The biggest hidden cost is custom metrics overage. Enterprise infrastructure hosts include only 200 custom metrics each. Any Kubernetes or microservices deployment easily generates thousands of unique timeseries, and overage adds up fast.

New Relic

New Relic charges per user and per GB of ingested data with no per-host fees. Annual pricing:

  • Full Platform Pro: $349/user/month
  • Core Users: $49/user/month
  • Basic Users: Free (unlimited)
  • Data Ingest: $0.40/GB beyond 100 GB free/month

The user-based pricing means your bill is heavily influenced by how many engineers need full platform access. Our estimates assume 20% Full Platform, 30% Core, and 50% Basic users, which is typical for engineering teams where not everyone needs query and alerting capabilities.

Grafana Cloud

Grafana Cloud Pro uses usage-based pricing across all signal types plus a per-user fee. The $19/month base platform includes 50 GB traces, 50 GB logs, 10k active series, and 3 users; overage beyond these thresholds is billed at:

  • Traces: $0.50/GB ingested
  • Logs: $0.50/GB ingested
  • Metrics: $6.50/1,000 active series
  • Users: $8/active user/month
  • Base platform: $19/month

The per-series metrics pricing can become the dominant cost for teams with high-cardinality metric workloads.

SigNoz

SigNoz Cloud charges per GB for traces and logs, and per million samples for metrics:

  • Traces: $0.30/GB
  • Logs: $0.30/GB
  • Metrics: $0.10/million samples
  • Base plan: $49/month (includes $49 of usage)

Metric sample counts depend on your collection interval. At a standard 1-minute interval, each active timeseries produces 43,200 samples per month.

Uptrace

Uptrace uses volume-tiered pricing for data (traces + logs combined) and active timeseries. Higher tiers unlock lower per-unit rates. Each tier sets a base price that covers a data allotment; timeseries are billed separately at the tier's per-series rate. Hetzner datacenter pricing:

  • Data: $0.05–$0.08/GB depending on tier (base covers an included allotment; overage at the same rate)
  • Timeseries: $0.0005–$0.0008/series/month depending on tier (billed on top of the data tier)
  • No per-host fees, no per-user fees, no indexing charges

The tiered model means per-unit costs decrease as you scale — the opposite of platforms where costs accelerate with volume.

Egress traffic note: Uptrace Cloud runs in Hetzner's Germany datacenter. If your infrastructure is on AWS or GCP, you pay standard cloud egress fees to send telemetry data to Hetzner. OTLP compresses data with protobuf + gzip (~5x compression), so the actual bytes on the wire are much smaller than the uncompressed volume Uptrace bills for. OTel Arrow reduces bandwidth by an additional 30–70% on top of that.

  • AWS egress: ~$0.09/GB (first 10 TB/month)
  • GCP egress: ~$0.11–0.12/GB (first 10 TB/month)

Detailed Cost Breakdown by Team Size

Large Team

150 engineers, 225 APM hosts, 350 infra hosts, 10 TB traces, 20 TB logs, 750k timeseries.

Datadog — $42,468/month

Line ItemCalculationCost
Infrastructure hosts350 × $23$8,050
APM hosts225 × $40$9,000
Log ingestion20,000 GB × $0.10$2,000
Log indexing (20% indexed)8,000M events × $1.70/M$13,600
Indexed spans (10% indexed)1,775M spans × $1.70/M$3,018
Custom metrics overage680k metrics × $0.01$6,800
Total$42,468

Log indexing assumes 20% of events indexed at 15-day retention. Average log event size of 500 bytes gives ~40 billion events from 20 TB. Indexed spans assume 10% of ~20 billion total spans, minus the 225M included with APM hosts.

New Relic — $24,635/month

Line ItemCalculationCost
Full Platform Pro (20%)30 users × $349$10,470
Core users (30%)45 users × $49$2,205
Basic users (50%)75 users × $0$0
Data ingest29,900 GB × $0.40$11,960
Total$24,635

User fees account for 51% of the total bill. Teams with more Full Platform users will see significantly higher costs.

Grafana Cloud — $20,955/month

Line ItemCalculationCost
Traces(10,000 − 50 incl.) GB × $0.50$4,975
Logs(20,000 − 50 incl.) GB × $0.50$9,975
Metrics(750 − 10 incl.) × $6.50 per 1k$4,810
Users(150 − 3 incl.) × $8$1,176
Base platform$19
Total$20,955

Log ingestion is the largest component at nearly half the total bill.

SigNoz — $12,240/month

Line ItemCalculationCost
Traces10,000 GB × $0.30$3,000
Logs20,000 GB × $0.30$6,000
Metrics32,400M samples × $0.10/M$3,240
Total$12,240

Metric samples calculated at 1 datapoint per timeseries per minute (750k × 43,200 samples/month).

Uptrace — $2,426/month

Line ItemCalculationCost
Data (traces + logs)30 TB at tier-5 rate ($0.05/GB)$1,511
Timeseries750k × $0.0005/series$375
AWS/GCP egress~6 TB compressed × $0.09/GB$540
Total$2,426

At tier 5, the $999 base covers 20 TB of data. The remaining ~10 TB of overage is charged at $0.05/GB. Timeseries are billed separately at the tier rate. Egress estimate assumes standard OTLP gzip compression (~5x), sending ~6 TB over the wire from 30 TB uncompressed. Using OTel Arrow would cut the egress cost roughly in half (~$270).

Midsize Team

65 engineers, 125 APM hosts, 200 infra hosts, 5 TB traces, 10 TB logs, 250k timeseries.

Datadog — $20,988/month

Line ItemCalculationCost
Infrastructure hosts200 × $23$4,600
APM hosts125 × $40$5,000
Log ingestion10,000 GB × $0.10$1,000
Log indexing (20% indexed)4,000M events × $1.70/M$6,800
Indexed spans (10% indexed)875M spans × $1.70/M$1,488
Custom metrics overage210k metrics × $0.01$2,100
Total$20,988

New Relic — $11,477/month

Line ItemCalculationCost
Full Platform Pro (20%)13 users × $349$4,537
Core users (30%)20 users × $49$980
Data ingest14,900 GB × $0.40$5,960
Total$11,477

Grafana Cloud — $9,525/month

Line ItemCalculationCost
Traces(5,000 − 50 incl.) GB × $0.50$2,475
Logs(10,000 − 50 incl.) GB × $0.50$4,975
Metrics(250 − 10 incl.) × $6.50 per 1k$1,560
Users(65 − 3 incl.) × $8$496
Base platform$19
Total$9,525

SigNoz — $5,580/month

Line ItemCalculationCost
Traces5,000 GB × $0.30$1,500
Logs10,000 GB × $0.30$3,000
Metrics10,800M samples × $0.10/M$1,080
Total$5,580

Uptrace — $1,349/month

Line ItemCalculationCost
Data (traces + logs)15 TB at tier-4 rate ($0.06/GB)$929
Timeseries250k × $0.0006/series$150
AWS/GCP egress~3 TB compressed × $0.09/GB$270
Total$1,349

At tier 4, the $499 base covers 8 TB of data. The ~7 TB overage is charged at $0.06/GB. Timeseries are billed separately at the tier rate. Egress estimate assumes ~5x OTLP compression on 15 TB uncompressed. OTel Arrow would reduce egress to ~$135.

Small Team

25 engineers, 20 APM hosts, 50 infra hosts, 1 TB traces, 2.5 TB logs, 75k timeseries.

Datadog — $4,856/month

Line ItemCalculationCost
Infrastructure hosts50 × $23$1,150
APM hosts20 × $40$800
Log ingestion2,500 GB × $0.10$250
Log indexing (20% indexed)1,000M events × $1.70/M$1,700
Indexed spans (10% indexed)180M spans × $1.70/M$306
Custom metrics overage65k metrics × $0.01$650
Total$4,856

New Relic — $3,497/month

Line ItemCalculationCost
Full Platform Pro (20%)5 users × $349$1,745
Core users (30%)8 users × $49$392
Data ingest3,400 GB × $0.40$1,360
Total$3,497

Grafana Cloud — $2,318/month

Line ItemCalculationCost
Traces(1,000 − 50 incl.) GB × $0.50$475
Logs(2,500 − 50 incl.) GB × $0.50$1,225
Metrics(75 − 10 incl.) × $6.50 per 1k$423
Users(25 − 3 incl.) × $8$176
Base platform$19
Total$2,318

SigNoz — $1,374/month

Line ItemCalculationCost
Traces1,000 GB × $0.30$300
Logs2,500 GB × $0.30$750
Metrics3,240M samples × $0.10/M$324
Total$1,374

Uptrace — $344/month

Line ItemCalculationCost
Data (traces + logs)3.5 TB at tier-3 rate ($0.065/GB)$232
Timeseries75k × $0.00065/series$49
AWS/GCP egress~700 GB compressed × $0.09/GB$63
Total$344

At tier 3, the $199 base covers 3 TB of data. The small overage is charged at $0.065/GB. Timeseries are billed separately at the tier rate. Egress estimate assumes ~5x OTLP compression on 3.5 TB uncompressed. OTel Arrow would reduce egress to ~$32.

What Drives Costs at Scale

Datadog: Log Indexing and Custom Metrics

Datadog's per-host fees are predictable, but log indexing and custom metrics create the steepest cost curves. In our large-team estimate, log indexing alone accounts for $13,600/month — and that assumes only 20% of logs are indexed. Teams that index a higher percentage or use longer retention will see this number grow rapidly.

Custom metrics overage is the other major driver. Each Enterprise infrastructure host includes only 200 custom metrics. A single Kubernetes cluster with standard exporters can generate tens of thousands of unique timeseries. At $0.01 per metric, 680k overage metrics cost $6,800/month.

New Relic: User Fees

New Relic's data ingest pricing ($0.40/GB) is straightforward, but user fees dominate the bill. Full Platform Pro users cost $349/month each. In our large-team estimate, user fees account for over half the total cost.

The key optimization lever is minimizing the number of Full Platform users. Moving engineers who only need dashboards to the free Basic tier can significantly reduce costs. However, teams where most engineers need query and alerting access will find New Relic expensive regardless of data volume.

Grafana Cloud: Metrics Series Pricing

Grafana Cloud's per-GB rates for traces and logs are competitive, but metrics pricing at $6.50 per 1,000 active series can surprise teams with high-cardinality workloads. For our large team, metrics account for $4,810/month.

Organizations that heavily rely on Prometheus-style metrics with many label combinations will see this cost grow faster than trace or log costs.

SigNoz: Metric Samples Volume

SigNoz offers simple per-GB pricing for traces and logs, but metric costs depend on your collection interval. At a 1-minute scrape interval, each timeseries produces 43,200 samples/month. Reducing the interval to 30 seconds doubles the sample count and the cost.

For our large team, metrics cost $3,240/month — a significant portion of the $12,240 total. Teams can optimize by increasing collection intervals for less critical metrics.

Uptrace: Predictable Scaling (Plus Egress)

Uptrace's volume-tiered model means per-unit costs decrease as you grow. There are no per-host fees, no per-user fees, and no separate indexing charges. The combined data rate (traces + logs billed together) simplifies budgeting.

At scale, the tier structure provides meaningful discounts: moving from tier 3 ($0.065/GB) to tier 5 ($0.05/GB) represents a 23% reduction in per-GB cost simply by committing to more volume.

The main additional cost to budget for is cloud egress if your infrastructure runs on AWS or GCP. Standard OTLP compression (~5x) keeps wire bytes manageable, and OTel Arrow reduces bandwidth by another 30–70%. For our large team, egress adds $540/month with standard OTLP or ~$270 with OTel Arrow — still a fraction of what competing tools charge for the data itself. Timeseries are billed separately on top of the data tier, adding $375/month for 750k series at the tier-5 rate.

Key Takeaways

Uptrace is the most cost-effective option at every team size, ranging from $344/month for a small team to $2,426/month for a large one (including AWS/GCP egress). The absence of per-host and per-user fees, combined with volume discounts, makes cost scaling predictable. Using OTel Arrow reduces the egress component by roughly half.

SigNoz offers the next-lowest costs with simple per-GB pricing and no host or user fees. However, metric sample-based pricing requires careful attention to collection intervals.

Grafana Cloud provides a balanced option for teams already invested in the Grafana ecosystem. Per-GB rates are moderate, but metrics series pricing can add up for high-cardinality workloads.

New Relic works well for teams with few Full Platform users and high data volumes. The user-based pricing model means small teams with proportionally more power users will pay more per GB of actual data.

Datadog offers the most comprehensive feature set but comes at the highest cost across all team sizes. The multi-dimensional pricing model (hosts + data + indexing + metrics) makes cost estimation and optimization complex.

FAQ

Are these estimates based on list pricing or negotiated rates?

All estimates use published list pricing with annual commitment where available. Enterprise customers typically negotiate discounts of 10–30%, especially at higher volumes. The relative ordering between tools is unlikely to change with standard discounts.

Why does the Datadog estimate seem lower than some published comparisons?

Our Datadog estimates assume only 20% of log events are indexed and 10% of spans are indexed. Teams that index a higher percentage — or use products like RUM, Synthetics, or CSPM — will see significantly higher totals. These estimates cover core observability only (APM, logs, metrics, infrastructure).

How do I reduce my New Relic bill?

The single most effective lever is reducing the number of Full Platform users. Move engineers who only need dashboard access to the free Basic tier. Also consider the Data Plus option ($0.60/GB) only if you need the additional features it provides, and stick with the standard $0.40/GB rate otherwise.

Does Grafana Cloud pricing include Grafana Enterprise plugins?

No. The estimates use Grafana Cloud Pro with standard visualization at $8/user. Grafana with Enterprise Plugins costs $55/user, which would increase the user component significantly.

How are SigNoz metric samples calculated?

One active timeseries scraped every 60 seconds produces 43,200 samples per month (60 minutes × 24 hours × 30 days). At $0.10 per million samples, that's about $0.004 per timeseries per month. Teams using 15-second scrape intervals would pay 4x more.

Why do Uptrace totals include egress costs but the others don't?

Uptrace Cloud runs in Hetzner's Germany datacenter. If your workloads are on AWS or GCP, your cloud provider charges standard internet egress fees to send telemetry out. We include this cost for transparency. The other tools in this comparison typically accept data at endpoints reachable over the internet as well — your cloud provider charges the same egress rate regardless of destination. The difference is that vendors hosted on the same cloud may offer VPC PrivateLink or peering options at enterprise tiers, which can reduce or eliminate egress. Uptrace currently does not offer PrivateLink.

Can Uptrace handle the same scale as Datadog or New Relic?

Yes. Uptrace uses ClickHouse for storage, which handles petabyte-scale workloads with 10–20x data compression. The Hetzner pricing shown here represents the managed cloud offering. AWS and Google Cloud deployments are also available at higher per-unit rates.