Observability pricing can look straightforward on a sales slide. The numbers seem reasonable, the team signs, and then the first full billing cycle lands higher than expected.

The problem is often that the pricing model wasn't stress-tested against how the team's usage would actually grow. Datadog and New Relic both offer a broad observability platform, and both will get your logs, metrics, and traces into a dashboard. Both are DevOps tools designed to help teams catch problems before customers do. But they charge for that data in genuinely different ways, and the difference compounds as your systems grow.

This post walks through how each vendor actually bills, not how they pitch it, so you can compare Datadog vs. New Relic pricing and size your own cost before you're locked into a contract.

How Datadog prices

Datadog sells observability as a set of separate products, and each one carries its own SKU. Teams can combine several of them for visibility across applications, infrastructure, logs, and user experience.

Here's how the core products are billed:

Infrastructure monitoring: priced per host. Around $15 a host per month on an annual commitment for the Pro plan, or $23 a host for Enterprise. On-demand rates run higher than the annual price.
Log management: split into two separate charges, log ingestion and log indexing. You pay a per-GB rate on ingest to bring logs in, then a separate per-million-events rate to make those logs searchable and queryable. A plan that looks affordable at the ingest rate can cost several times more once indexing is added for the logs your team actually searches.
Synthetic monitoring and real user monitoring: billed on their own units too, stacked on top of infrastructure and APM.

Then there's the container ceiling. Infrastructure Monitoring includes 5 containers per host on Pro and 10 on Enterprise, with additional containers billed separately. That matters more than it sounds like for teams running Kubernetes, where container density per host tends to climb over time, not shrink. Container limits and the gap between on-demand and annual-commit pricing, sometimes 17% to 55% apart depending on the product, are two of themost common observability pricing traps, and both are worth asking about directly before you sign anything.

How New Relic prices

New Relic prices on two main variables:

How much data you ingest.
The type and number of users who need access to the platform.

Data ingest:

Billed per GB, with 100 GB free every month across all data types, including logs, metrics, traces, and events.
Data can arrive through New Relic's own agents or through OpenTelemetry, billed the same either way.
Data Plus, an optional add-on, layers on extended data retention and additional compliance and export features at a slightly higher per-GB rate.

Users:

Basic users: free, unlimited, with dashboard and alerting access.
Core users: $49 a user per month, adding error tracking, log access, and deeper IDE-level telemetry on top of basic access.
Full platform users: unlock the complete toolset, including application performance monitoring, infrastructure monitoring, distributed tracing, real user monitoring and synthetic monitoring (together sometimes called digital experience monitoring), session replay, custom metrics, and querying every bit of that data through NRQL (New Relic Query Language) against NRDB, the underlying data store. On New Relic's free tier, the first full platform user is included at no cost — paid editions (Standard, Pro, Enterprise) charge for every full platform user, including the first.

The practical difference from Datadog's model: New Relic's pricing units don't multiply as you add capabilities. Turning on database monitoring, network performance monitoring, error tracking, or serverless monitoring doesn't add a new SKU. It's already included for anyone with full platform access, and it pulls from the same ingest pool.

If you're comparing the two models directly,New Relic's Datadog comparison breaks down what's included with each platform and where separate charges apply. The structural point is simple: fewer units to track can make costs easier to forecast as your usage changes.

For the numbers behind that model, New Relic's current pricing shows the latest per-GB and per-user rates. Rate cards shift over time on both sides, so those figures are the ones to use when you model your own costs.

What actually drives your bill

Look beyond the headline prices, and both models respond to the same underlying pressures. What changes is which one you feel first.

Data volume

The biggest lever in either case. Applications built on microservices and containers can generate much more telemetry than monolithic applications, simply because there are more services talking to each other and more places to instrument.

The 2025 CNCF Annual Cloud Native Survey found that 82% of container users now run Kubernetes in production, up from 66% just two years earlier. As architectures become more distributed, teams have more potential sources of distributed tracing and log data to account for in their observability budget.

Host and container count

Host and container growth has a more direct effect on Datadog's model, since infrastructure and APM pricing scale directly with host count. A team that grows its footprint through auto-scaling or a Kubernetes migration can see its bill move even if headcount stays flat.

Team size

This matters too, but less than most buyers assume. Data can grow much faster than headcount. A five-person platform team can generate an enormous amount of telemetry from a large fleet of services, so team size alone doesn't tell you much about observability usage. A pricing model anchored to data volume instead tracks the amount of telemetry you're actually sending.

Add-ons

The final lever is what sits outside the core pricing model. For New Relic, that can include Data Plus for capabilities such as extended data retention, as well as pay-per-use Advanced Compute, which is measured in Advanced Compute Capacity Units (aCCUs). New Relic's AI capabilities include SRE Agent for automated root cause analysis.

Datadog has its own additional pricing units to consider. Bits AI SRE is priced through AI credits, while capabilities such as security monitoring and cloud security posture management sit within separately priced security products.

Datadog's Watchdog provides AI-powered anomaly detection and root cause analysis within the platform. None of this is wrong on its own, but each additional pricing unit is something else to account for when forecasting next year's bill.

Total cost of ownership at different team sizes

Sticker price tells you only part of the story, because the same feature set can produce very different bills depending on how your organization is shaped.

Small team

A small team running a handful of services may have relatively few hosts, users, and GB of data to account for. New Relic's free tier can also absorb some early usage, while Datadog's cost depends on which products the team needs and the infrastructure running them.

At this stage, model what happens if your data volume, host count, or product usage doubles. The cheapest starting price isn't necessarily the pricing model that will stay cheapest as you grow.

Mid-size engineering org

Once a team is running a larger microservices footprint, more of the pricing levers start moving at once. Host count and container density climb, log volume grows with the number of services in play, and more engineers may need access to observability data.

For Datadog, that means modeling the hosts, containers, data, and separately priced products you'll use. For New Relic, the main variables remain data ingest and user type. DevOps teams should also account for the capabilities they'll actually need rather than comparing entry-level plans alone. Aninfrastructure monitoring evaluation can help you identify those requirements before you start comparing costs.

Large enterprise

With hundreds of services and a mature observability practice, you'll usually negotiate custom terms with either vendor, so public rate cards matter less than the underlying model. What still matters is how predictable the number is year over year. Worldwide IT spending is projected to grow another 14.2% in 2026 according to Gartner, with software spending growing even faster than that overall figure. Against that backdrop, predictable observability costs become increasingly useful for annual planning.

Run realistic scenarios for the variables most likely to change:

Data ingest: How much more telemetry will you send as your environment grows?
Hosts and containers: How will auto-scaling or infrastructure growth affect host-based charges?
Full platform users: How many people will need deeper access as your DevOps teams grow?
Data retention: Will you need to keep telemetry for longer?
Additional capabilities: Will features or add-ons introduce another pricing unit?

That's where the difference between Datadog's product-specific pricing units and New Relic's data-and-user model becomes more useful than a simple comparison of list prices.

Dollar figures aren't pinned down here on purpose. Rate cards can change, so always refer to:

Datadog's current pricing
New Relic's current pricing

Run your own usage numbers through each model before you compare totals.

Questions to ask before you commit

A short checklist to walk through with any vendor, not just these two:

How does billing handle traffic spikes or auto-scaling? Ask whether charges are based on average, peak, or another measure of usage, and what happens when infrastructure scales up temporarily. A short-lived traffic spike can affect different pricing models in very different ways.
Is pricing per-host or usage-based? Per-host pricing can increase costs as you scale horizontally. Usage-based pricing tracks more closely with what you're actually sending, but ask specifically what counts as "usage," since the definition varies by vendor.
What's bundled versus billed separately? Get a straight answer on whether distributed tracing, real user monitoring, synthetic monitoring, log management, and anomaly detection ship with the base plan or require an additional SKU. If you're also evaluating alternatives beyond these two,this rundown of Datadog alternatives compares how other observability platforms package and price their capabilities.
Do overage fees exist, and what triggers them? Ask for the exact GB, host, or container ceiling on your plan and what happens the moment you cross it. Make sure you know both the threshold and the rate you'll pay beyond it before you commit.

Choosing the pricing model that fits your team

Sticker price on either vendor's homepage doesn't tell you much on its own. What matters is how each pricing model behaves once your data grows, your team scales, and your architecture gets more distributed than it is today. A model built around separate product SKUs can introduce additional billing units as you add capabilities. A model built around two main units, data ingest and full platform users, keeps the variables you need to forecast more consistent as you scale.

New Relic's usage-based pricing is built around that predictability: one ingest number, one user count, a free tier that covers real usage for smaller teams, and a clear path to scale without adding new product lines to your invoice.

If you want to model what New Relic would look like for your own environment,request a demo to see the platform with your data and requirements.