About 18; min
Datadog’s pricing page lists clean per-host numbers that look reasonable. Then you enable APM, add log ingestion, turn on Synthetics, and suddenly your monthly bill rivals a junior engineer’s salary. We’ve seen this happen to dozens of teams.
Here’s what Datadog actually costs in 2026, broken down by product, with real scenarios so you can budget before you get surprised.
Datadog Product Pricing at a Glance
| Product | Free Tier | Pro | Enterprise |
|---|---|---|---|
| Infrastructure Monitoring | 5 hosts | $15/host/mo | $23/host/mo |
| APM | — | $31/host/mo | $40/host/mo |
| Log Management (ingest) | — | $0.10/GB/mo | $0.10/GB/mo |
| Log Management (retain 15d) | — | $1.70/million events | $1.70/million events |
| Synthetics (API tests) | — | $5/10K runs/mo | $12/10K runs/mo |
| Synthetics (Browser tests) | — | $12/1K runs/mo | $18/1K runs/mo |
| RUM | — | $1.50/1K sessions | $1.80/1K sessions |
| CSPM | — | $7.50/host/mo | $11/host/mo |
These are list prices. Committed-use contracts (annual, prepaid) typically save 15–25%, but require you to predict usage accurately — and overages get billed at on-demand rates.
Infrastructure Monitoring: The Foundation
The free tier gives you 5 hosts with 1-day retention. That’s enough to test Datadog on a staging environment but useless for production monitoring. Most teams land on Pro at $15/host.
The Pro-to-Enterprise jump adds AI-powered anomaly detection (Watchdog). If you have 200+ hosts, Watchdog can surface issues faster than manual alerting rules. Below 50 hosts, most teams find Pro’s standard monitors sufficient.
The Container Factor
Datadog counts containers differently from hosts. You get 10 containers free per host in the Pro plan and 20 in Enterprise. Beyond that, each additional container costs $0.002/hour (~$1.44/month). For a Kubernetes cluster running 500 pods across 20 nodes, you’d pay for the 20 hosts plus roughly 300 excess containers = $432/month extra. That adds up fast in microservice architectures.
APM: Where Costs Get Serious
APM is priced per host, not per service. A single host running 12 microservices still counts as one host for billing. However, each host includes 1 million Indexed Spans per month. If your services generate more traces than that, additional spans cost $1.70 per million.
Here’s where it gets expensive: a medium-traffic API service on a single host can easily generate 5–10 million spans per month. That’s $6.80–$15.30 in overage per host — adding 20–50% on top of the base $31.
Controlling APM Costs
Datadog offers Ingestion Controls to set sampling rates per service. Most teams find that sampling 10–20% of traces gives them enough data for debugging while cutting span costs by 80%. The tricky part: you need to configure this before you start paying overage, not after.
Log Management: The Budget Killer
Log management is where most Datadog sticker shock happens. The pricing has three components:
- Ingestion: $0.10/GB — every byte you send to Datadog
- Indexing (15-day retention): $1.70/million log events — only indexed logs are searchable
- Archival/Rehydration: $0.05/GB for S3 archive, $0.10/GB to rehydrate
A mid-size application generating 100 GB/day of logs costs:
| Component | Monthly Cost |
|---|---|
| Ingestion (100 GB/day × 30) | $300 |
| Indexing (~50M events, 10% sampled) | $85 |
| Archive to S3 | $150 |
| Total | $535/mo |
That’s one application. A company running 20 services at similar log volumes is looking at $10,000+/month just for logs. This is why many teams use Datadog for metrics and APM but ship logs to a cheaper alternative like Grafana Loki or self-hosted Elasticsearch.
Log Pipelines and Exclusion Filters
Datadog lets you create exclusion filters that drop specific log patterns before indexing. Teams that aggressively filter health checks, debug-level logs, and repetitive warnings typically cut log costs by 60–70%. Setting up these filters should be your first action after enabling log management.
Real Cost Scenarios
Startup (10 hosts, basic monitoring)
| Product | Monthly Cost |
|---|---|
| Infrastructure Pro (10 hosts) | $150 |
| APM Pro (5 API hosts) | $155 |
| Logs (10 GB/day) | $55 |
| Total | $360/mo ($4,320/yr) |
Mid-size (50 hosts, full stack)
| Product | Monthly Cost |
|---|---|
| Infrastructure Pro (50 hosts) | $750 |
| APM Pro (30 hosts) | $930 |
| Logs (50 GB/day) | $285 |
| Synthetics (50 API + 20 browser) | $49 |
| RUM (100K sessions) | $150 |
| Total | $2,164/mo ($25,968/yr) |
Enterprise (200 hosts, full observability)
| Product | Monthly Cost |
|---|---|
| Infrastructure Enterprise (200 hosts) | $4,600 |
| APM Enterprise (100 hosts) | $4,000 |
| Logs (500 GB/day) | $2,200 |
| Synthetics (200 API + 100 browser) | $220 |
| RUM (500K sessions) | $750 |
| CSPM (200 hosts) | $2,200 |
| Total | $13,970/mo ($167,640/yr) |
At the enterprise level, Datadog costs more than many team members. This is why committed-use contracts matter: a 3-year commitment at this scale typically saves 25–30%, bringing the annual cost down to roughly $117,000–$126,000.
Datadog vs New Relic vs Grafana Cloud
| Feature | Datadog | New Relic | Grafana Cloud |
|---|---|---|---|
| Pricing model | Per-host + usage | Per-user + GB ingested | Usage-based |
| Free tier | 5 hosts, 1-day retention | 100 GB/mo free, 1 user | 10K metrics, 50 GB logs |
| Infrastructure (50 hosts) | $750/mo | ~$500/mo (depends on users) | ~$350/mo |
| APM | $31–$40/host | Included in GB pricing | $20/host + traces |
| Logs (50 GB/day) | ~$285/mo | ~$150/mo (at scale) | ~$120/mo |
| Dashboard quality | Excellent | Good | Excellent (Grafana) |
| AI features | Watchdog, Bits AI | AI monitoring | Limited |
| Setup complexity | Low (agent-based) | Low (agent-based) | Medium (more DIY) |
New Relic is often cheaper for smaller teams because its per-user model means you don’t pay per host. But costs scale unpredictably with data volume. Grafana Cloud is the budget option — open-source Grafana, Loki, Tempo, and Mimir managed for you at roughly 40–50% less than Datadog. The trade-off is more setup work and fewer out-of-the-box integrations.
Cost Optimization Tips
- Set sampling rates early: Configure APM trace sampling and log exclusion filters before your first month closes
- Use custom metrics wisely: Each host includes 100 custom metrics in Pro. Going over costs $0.05/metric/month. Tag carefully
- Right-size your commitment: Datadog offers 1-year and 3-year committed-use discounts. Run for 2–3 months on-demand first to understand your actual usage patterns
- Separate environments: Don’t send staging/dev metrics to your production Datadog org unless you need them. Each host counts
- Archive aggressively: Move logs to S3 after 3 days instead of keeping 15-day indexed retention for everything
Best for: Teams that want a single platform for metrics, traces, logs, and synthetics with minimal setup. Datadog’s agent installs in minutes and the 500+ integrations mean you’re usually monitoring within an hour.
Watch out for: Log costs. They will surprise you. Budget 2–3x your initial estimate for the first quarter until you tune exclusion filters and sampling rates.
Consider alternatives if: Your budget is under $1,000/month — New Relic or Grafana Cloud will give you more per dollar at that scale.
FAQ
Does Datadog charge for containers?
You get 10 free containers per host (Pro) or 20 (Enterprise). Beyond that, each container costs $0.002/hour. Serverless functions have separate pricing at $2/million invocations.
Can I use Datadog for just infrastructure monitoring?
Yes. You can enable products individually. Many teams start with Infrastructure Pro only and add APM or logs later as needed.
How do committed-use contracts work?
You prepay for a set number of host-months and GB at a discounted rate. If you exceed your commitment, overages are billed at on-demand rates. Unused capacity doesn’t roll over.
Is the free tier useful for anything?
It’s fine for a small staging environment or a side project with 5 or fewer servers. The 1-day metric retention makes it impractical for production troubleshooting.
