CompareUpdated 2026-09-22
APM pricing models: Datadog vs Dynatrace vs New Relic
Datadog charges per host with separate meters for logs, custom metrics and each product, Dynatrace draws every capability down from one annual commitment and charges full-stack hosts by their memory, and New Relic charges per GB of data ingested plus either users or compute, with no host count at all.
18cells, every one sourced and dated
Side by side
From the vendors' own pages
- Datadog pricing list
- Dynatrace pricing
- New Relic pricing
- Datadog custom metrics billing
- Datadog OpenTelemetry docs
- New Relic OTLP docs
A cell without a source does not ship.
01The mechanics
What actually differs
The unit each option charges on, and who owns what afterwards, matter more than any single quoted figure.
APM (application performance monitoring) watches how your software behaves in production: how long each request takes, where it slows down and why it fails. Its pricing is where bills surprise hardest, because the meters multiply. At Datadog, for example, an application host is billed on top of the infrastructure host it runs on, and a log is charged once to ingest and again to index. Each vendor's model makes a different habit expensive, so the same estate can cost very different amounts depending on which meter its habits land on.
The table below reads each vendor's own pricing page, checked on 22 September 2026. These are list prices, and contracts discount them, so read the numbers as the shape of each model rather than a quote. What matters most is the unit: per host, per GiB (roughly a gigabyte) of host memory, or per GB of data and per person. The last row covers OpenTelemetry, the open standard for collecting this data, because it decides how easily the data can later go somewhere else.
02The facts
Side by side, sourced
Every cell in this table carries the vendor's own page as its source and the date it was checked, and a cell without a source does not ship.
| Datadog | Dynatrace | New Relic | |
|---|---|---|---|
| What you pay on | Per host per month, with a separate price for each product and for logs and custom metrics.Datadog Infrastructure Pro is $15 per host a month billed annually, $18 on demand.Datadog pricing list, checked 2026-09-22 | One annual spend commitment for the whole platform, drawn down at published rate-card prices per host-hour, per GiB of host memory per hour, or per GiB of data.Dynatrace Infrastructure Monitoring is $29 per host a month, billed at $0.04 per host per hour.Dynatrace pricing, checked 2026-09-22 | Per GB of data ingested, plus either a price per user or a compute price that covers every user. Hosts are not counted.New Relic includes 100 GB of ingest a month, then charges $0.40 per GB.New Relic pricing, checked 2026-09-22 |
| Application monitoring on one host | Billed per APM host, on top of the infrastructure host.Datadog APM is $31 per APM host a month billed annually, $36 on demand. APM Pro is $35 and APM Enterprise $40, billed annually.Datadog pricing list, checked 2026-09-22 | Billed by host memory, so a bigger host costs more.Dynatrace Full-Stack Monitoring is $58 a month for an 8 GiB host, billed at $0.01 per GiB of memory per hour. It includes infrastructure monitoring.Dynatrace pricing, checked 2026-09-22 | No per-host charge. New Relic includes unlimited hosts, agents and containers, and bills the data they send.New Relic pricing, checked 2026-09-22 |
| Logs | Charged to ingest, then charged again for each event you index to search.Datadog log ingest is $0.10 per GB. Indexing is $1.06 to $2.50 per million events a month billed annually, rising with 3 to 30 days of retention.Datadog pricing list, checked 2026-09-22 | Charged to ingest, then for each day kept, with queries either charged per GiB scanned or included in a higher retention rate.Dynatrace log ingest is $0.20 per GiB. Retention is $0.0007 per GiB-day with queries at $0.0035 per GiB scanned, or $0.02 per GiB-day with queries included.Dynatrace pricing, checked 2026-09-22 | The same per-GB ingest rate as every other kind of data, and all of it indexed and ready to query.New Relic Data Plus is $0.60 per GB and adds up to 90 days of extra retention for most data types.New Relic pricing, checked 2026-09-22 |
| Custom metrics | An allotment per host, pooled across the account, and anything above it priced in your contract.Datadog includes 100 custom metrics per host on Pro and 200 on Enterprise. Each unique combination of metric name and tag values counts as one.Datadog custom metrics billing, checked 2026-09-22 | An allowance included with Infrastructure and Full-Stack Monitoring, then charged per datapoint.Dynatrace metrics ingest is $0.15 per 100,000 datapoints, with 15 months of retention included and no charge to query.Dynatrace pricing, checked 2026-09-22 | No separate metric meter. Metrics are billed as data ingested, like everything else.New Relic pricing, checked 2026-09-22 |
| Commitment | Billed annually, month to month, or on demand. The annual rate is the lowest.Datadog APM is $31 per host a month billed annually and $36 month to month.Datadog pricing list, checked 2026-09-22 | A minimum annual spend at the platform level. Usage above it is billed monthly at the same rates, with no penalty overage.Dynatrace pricing, checked 2026-09-22 | Pay as you go monthly on Standard and Pro, cancel at any time, or an annual or multi-year savings plan on Pro and Enterprise.New Relic rolls unused savings-plan funds over for up to 12 months.New Relic pricing, checked 2026-09-22 |
| OpenTelemetry | Accepted three ways: the Datadog Agent with its own OpenTelemetry Collector build (Datadog's recommended path), a standard OpenTelemetry Collector, or direct ingest over OTLP (the OpenTelemetry Protocol).Datadog OpenTelemetry docs, checked 2026-09-22 | Native OTLP ingest for metrics and traces, priced at the same rates as Dynatrace's own data.Dynatrace pricing, checked 2026-09-22 | Native OTLP ingest for traces, metrics and logs, which New Relic names its preferred way to receive OpenTelemetry data.New Relic OTLP docs, checked 2026-09-22 |
03The honest row
When each is the right answer
Every option in this comparison is the right answer for somebody, and saying when is the part most comparisons leave out.
Datadog
The right answer when developers already work in it and the estate is cloud-native. A flat rate per APM host is kind to large-memory hosts, and the trace views earn their price on the services that matter most. The log indexing and custom-metric meters are where governing what arrives pays back fastest.
Dynatrace
The right answer for large, deeply distributed enterprise applications where automated root cause analysis shortens incidents, and where one annual commitment across every capability suits how the business buys. Because Full-Stack Monitoring is charged by host memory, estates of many small hosts are its best case.
New Relic
The right answer when the estate has many hosts and a small group of engineers who need the full platform, because hosts are not counted and the bill follows data and people. Monthly pay as you go on Standard and Pro suits a team that will not sign an annual commitment.
04Questions
Asked about this comparison
Why is APM spend so hard to predict?
Because the meters compound: hosts, custom metrics, log ingest and indexing, synthetic tests, and per-product add-ons each bill separately. Estates rarely watch every meter at once, so growth lands quietly on whichever one nobody owns. One metering view by source and destination is the fix that comes before any vendor change.
Why does the same host cost different amounts at each vendor?
Because each vendor counts something different. Datadog counts the host, whatever its size. Dynatrace Full-Stack Monitoring counts the host's memory, so a 32 GiB host costs four times an 8 GiB one. New Relic does not count the host at all and bills the data it sends plus the people who look at it.
Is switching APM vendors worth it for price alone?
Rarely. Instrumentation is sticky and engineers' attachment to their workflow is real. The moves that pay reliably are governing what reaches the meters and taking ordinary infrastructure hosts off APM-grade pricing, and both work whichever vendor you keep.
Model it against your estate
The comparison that matters is the one run on your volumes and your contracts. The review reads them with you, and you leave with your version of the Logmetry Blueprint. No system access, no obligation.
Free. No system access. No obligation. You keep the drawing.
What you get
- Your architecture today, drawn
- The waste map
- Costs before and after, with estimates
- Your target architecture
- Options for your business problem
What it costsNothing