Datadog vs New Relic: Pricing Model Explained
| Dimension | Datadog | New Relic |
|---|---|---|
| Infrastructure monitoring | $15/host/mo (Pro), $23/host/mo (Enterprise) | Included in data ingest pricing |
| APM (application tracing) | $31/host/mo (Pro), $40/host/mo (Enterprise) | Included in data ingest pricing |
| Logs | $0.10/GB ingested + $0.05/GB/mo retention (15d) | $0.30/GB ingest (all data incl. logs, metrics, traces) |
| Users | Included (no per-user charge) | Free users: unlimited. Full Platform users: $99–$549/mo |
| Metrics | $0.05/custom metric/mo | Included in ingest |
| Synthetics | $5/1,000 test runs/mo | $0.005/check run/mo |
| Dashboards | Included | Included |
| Alerts | Included | Included |
| Security monitoring | $0.20/GB add-on | Included with data |
| On-call / incident mgmt | $19/user/mo add-on | Included in Full Platform |
| Data retention | 15 days default (+$0.05/GB/mo for 30d) | 8 days (30–90d upgrade available) |
Both vendors offer committed spend discounts. Datadog typically offers 15–30% off list for annual commits above $50K. New Relic offers volume-tiered pricing below $0.30/GB at scale (above 500GB/mo ingest).
Real-World Cost Scenario: 50-Host SaaS Company
| Component | Datadog Cost | New Relic Cost | Grafana Cloud |
|---|---|---|---|
| Infrastructure (50 hosts) | $750/mo (Pro) | ~included in ingest | $8/host/mo = $400/mo |
| APM (50 services traced) | $1,550/mo | ~included in ingest | $0.10/GB traces |
| Logs (500 GB/mo ingest) | $50/mo ingest + $75/mo retention | $150/mo (500GB × $0.30) | $0.50/GB = $250/mo |
| Custom metrics (500 metrics) | $25/mo | ~included | Included |
| Synthetics (100K runs/mo) | $500/mo | $500/mo | $200/mo |
| Users (20 engineers) | Included | $1,980/mo (20 × $99 Full Platform) | Included |
| Estimated Monthly | $2,950/mo ($35,400/yr) | $2,630/mo ($31,560/yr) — with full-platform users | $850–$1,200/mo ($10,200–$14,400/yr) |
New Relic scenario assumes Full Platform users for all 20 engineers — cheaper if most users are "basic users" who only view dashboards (Basic users are free). Grafana Cloud scenario uses Prometheus+Loki+Tempo OSS stack with Grafana Cloud hosting.
Why Datadog Bills Spiral
Log Volume Surprises
Bills spike with incidents
During a production incident, log volume can spike 5–20× normal. Datadog charges per GB ingested — a 4-hour incident with high log volume can add $500–$5,000 to your monthly bill. Teams often discover the cost of debugging their most critical incidents.
Custom Metrics Tax
$0.05/metric/mo
Every custom business metric costs $0.05/mo. Active monitoring with Prometheus can generate 10,000–100,000 custom metrics easily. At 50,000 metrics, that's $2,500/mo = $30,000/yr just for custom metrics — often discovered at first renewal when the team has "instrumented everything."
APM Host Creep
Every host you add costs $31/mo
Datadog APM charges per host running an agent — including ephemeral containers. Auto-scaling Kubernetes clusters with burst capacity can double your host count during traffic peaks. Datadog uses the 99th percentile host count over the month, meaning burst events affect your bill.
Product Add-On Sprawl
$10–$20/host/mo each
Datadog's product portfolio is extensive: Network Performance ($8/host/mo), Cloud Cost Management ($0.025/container hour), Security (+$0.20/GB), Database Monitoring ($70/host/mo), CI Visibility ($0.004/CI pipeline minute). Each looks cheap individually but compounds quickly.
Retention Tiers
$0.05/GB/mo for 30-day logs
Default log retention is 15 days. Compliance or SLA requirements often need 30, 60, or 90 days. Each extension charges per GB per month of additional retention. A team with 1TB/mo of logs paying for 30-day retention adds $50/mo in retention fees — $600/yr just for the extra 15 days.
Committing Wrong
Annual commit = locked in
Datadog annual contracts lock in committed spend. If you scale down (layoffs, architecture changes), you still pay the committed amount. New Relic's data-based pricing adjusts more naturally with scale — you only pay for what you ingest, with committed spend discounts separate from usage.
When Each Platform Wins
Choose Datadog when:
Best for large, stable infra
You have predictable host counts (not heavily auto-scaling). You need the broadest ecosystem integrations — Datadog's 650+ integrations remain best-in-class. You're an AWS, GCP, or Azure power user and benefit from cloud provider discounts (Datadog has marketplace offerings that count toward cloud committed spend). Your security team needs Datadog's SIEM and security platform. You're above $200K/yr and can negotiate committed spend deals (25–35% discounts at scale).
Choose New Relic when:
Best for predictability and most engineers
You want predictable bills that don't spike with incidents. Most of your engineers only view dashboards (Basic users are free — only Full Platform users at $99–$549/mo). You ingest lower log volumes (under 200GB/mo) where New Relic's ingest model is cheaper per signal type. You want APM + infrastructure + logs + synthetics in one bill without the per-product billing complexity of Datadog.
Choose Grafana Cloud when:
Best for cost-conscious engineering teams
You have Prometheus/OpenTelemetry already set up (or are willing to). You want 70–85% savings vs Datadog for a Metrics + Logs + Traces stack. Your team is comfortable with OSS tooling (Grafana dashboard UX, PromQL, LogQL). You don't need enterprise support SLAs. Series A–B startups and engineering-led companies consistently prefer Grafana for the cost-to-capability ratio. Grafana Cloud free tier is genuinely useful (50GB logs, 10K metrics, 14-day retention free).
Consider Dynatrace when:
Enterprise, AI-driven
You need AI-powered root cause analysis (Dynatrace Davis AI). Your environment is complex (hybrid cloud, microservices, mainframe). You're willing to pay 20–30% more than Datadog for automated problem detection and full-stack causation. Dynatrace pricing ($69/host/mo for Full Stack) is higher list but includes capabilities that Datadog charges as add-ons (code-level tracing, user sessions, business analytics).
Observability Platform Comparison: Quick Reference
| Platform | Pricing Model | Est. Cost (50 hosts) | Best Strength |
|---|---|---|---|
| Datadog | Per host + per GB + per product | $35K–$80K/yr | Broadest integrations, best UX |
| New Relic | Per GB ingest + free basic users | $25K–$50K/yr | Predictable pricing, free users |
| Dynatrace | Per host (full-stack $69/host/mo) | $40K–$120K/yr | AI root cause, full-stack auto |
| Grafana Cloud | Per GB metrics/logs/traces | $8K–$18K/yr | 70–85% cheaper, OSS stack |
| Elastic Observability | Per GB ingest | $12K–$30K/yr | Log-heavy workloads, ELK ecosystem |
| AWS CloudWatch | Per metric + per GB logs | $5K–$25K/yr (AWS-only) | AWS-native, zero setup for AWS |
| VictoriaMetrics (OSS) | Self-hosted (infra cost only) | $1K–$5K/yr | Cheapest for metrics-heavy; ops overhead |
5 Cost Reduction Tactics if You Stay on Datadog
1. Log Sampling with Observability Pipelines
40–60% log cost reduction
Use Cribl Stream or Datadog's own Observability Pipelines (formerly Vector) to sample debug/info logs before they hit Datadog. Keep 100% of ERROR/CRITICAL logs, sample 10% of INFO logs, drop DEBUG entirely. Average result: 40–60% log volume reduction without losing signal fidelity. Command:
cribl pack logs --filter level=DEBUG --sampling 0.05.2. Custom Metric Audit
$5K–$30K/yr savings
Run a custom metrics audit monthly:
datadog-api-client metrics list --active. Unused custom metrics from decommissioned services or dead experiments continue billing. Typical finding: 20–40% of custom metrics haven't been queried in 30+ days. Remove agents from unused hosts and prune metric submissions from dead code paths.3. Move to Annual Committed Spend
15–30% discount
Datadog's list prices are month-to-month rates. Annual commits at $50K+ get 15–20% off. Annual commits at $200K+ get 25–30% off. Always negotiate at end of Datadog's fiscal quarters. If you're up for renewal, get competitive quotes from New Relic and Grafana Cloud — Datadog AEs have significant discretionary discount authority when facing competitive pressure.
4. Right-Size APM Tracing (Tail Sampling)
$8K–$25K/yr savings
Datadog APM charges per host but also charges for retained traces above free tier. Enable tail-based sampling to keep 100% of error traces and 1% of success traces. Trace volume reductions of 80–95% are achievable without losing error debugging capability. Configure via
DD_TRACE_SAMPLE_RATE=0.01 with error trace override.5. Consolidate Log Pipelines
30–50% log cost reduction
Many teams double-send logs: CloudWatch → Datadog (lambda forwarder) AND direct agent → Datadog. Audit your log ingestion paths — duplicate sources are common after migrations. Also: set log exclusion filters in Datadog for health-check endpoints, static file requests, and load balancer probes that create volume with zero debugging value.
Case Studies
Series B SaaS (80 hosts)
Infrastructure + APM + Logs
$48K/yr
Datadog Enterprise with APM + logs: $8,200/mo ($98,400/yr). Audit found: 22 hosts running Datadog agent but doing no work (dev instances, failed decommissions), custom metrics at 85,000 (used ~12,000 actively), logs at 2.2TB/mo with 70% being access logs and health-check endpoints. After cleanup: removed 22 hosts, pruned metrics to 15K, added Cribl log sampling (reduced to 650GB/mo). New bill: $4,200/mo ($50,400/yr). Savings: $48K/yr.
DevOps Consultancy (30 clients)
Multi-tenant observability
$96K/yr
Was running Datadog for 30 client environments at $4,500/mo average = $135K/mo ($1.62M/yr). Migrated to Grafana Cloud with per-client Prometheus + Loki stacks. Setup time: 8 weeks. Client environments now cost average $800/mo in Grafana Cloud = $24K/mo ($288K/yr). Savings: $111K/mo ($1.33M/yr) — shared with clients as "observability efficiency savings." Retained all 30 clients. Consultant's gross margin on observability services improved from 22% to 61%.
FinTech Platform (120 hosts)
High log volume, compliance
$180K/yr
Datadog at $18K/mo ($216K/yr) — driven primarily by log volume (4TB/mo at $0.10/GB + $0.05/GB retention for 90d). Compliance required 90-day retention. Negotiated with Datadog for 20% committed discount ($172,800/yr). Then migrated logs to Elastic Cloud ($0.05/GB ingest, 90d retention included) keeping Datadog for metrics/APM only. Log reduction strategy: dropped DEBUG and health-check logs via Cribl (70% log volume reduction). New combined bill: Datadog $80K/yr + Elastic $16K/yr = $96K/yr. Net savings: $120K/yr vs original. Vs negotiated rate: saved $77K/yr. Total documented savings over 12 months vs original rate: $180K.
Know When Datadog Raises Your Price
PricePulse tracks Datadog, New Relic, Grafana, Dynatrace pricing and alerts you before renewal — giving you leverage to negotiate or evaluate alternatives.
Start Free Observability Cost Audit →