New Relic vs Datadog
Both platforms do APM, infrastructure, logs, RUM and synthetics, and neither can be self-hosted. The comparison is not really about features, it is about which noun you are billed for. New Relic charges for the engineers who use it. Datadog charges for the hosts you monitor and for each module you switch on. Your own ratio of people to hosts settles the argument before any feature list does.
Pick Datadog if you have more engineers than hosts: seats are free, so a twenty-person team on twenty hosts pays $920 a month for infrastructure and APM rather than $6,980 in New Relic Pro seats. Stay on New Relic if you have more hosts than engineers: five or fewer full platform users cost at most $406 a month regardless of how many containers, functions and nodes report in, and 100 GB of ingest is free every month.
Side by side
New Relic vs Datadog at a glance
| New Relic | Datadog | |
|---|---|---|
| Bills by | Seats, GB ingested and compute units | Hosts, modules, GB and indexed events |
| Full access per seat | ✗ $349/user on Pro, $549 on Enterprise | ✓ no seat fee on infrastructure or APM |
| Seat ceiling on the cheap tier | ✗ Standard stops at 5 full platform users | ✓ unlimited readers |
| Per-host cost | ✓ none, hosts are not a billing unit | ✗ $15/host Infra Pro, $23 Enterprise, plus $31/host APM |
| Free plan | ✓ 100 GB/mo, 1 full user, 8-day retention | ✗ 5 hosts, 1-day metric retention |
| Log ingest | $0.40/GB, or $0.60/GB Data Plus | ✓ $0.10 per ingested or scanned GB |
| Log search | Included in the ingest price | ✗ indexing billed separately, $1.70 per million events |
| Custom metrics | Counted as ingested GB | ✗ 100 per host included, then billed per series |
| Query language | ✗ NRQL, portable nowhere | ✗ proprietary query and monitor syntax |
| Self-host / licence | ✗ none, SaaS only | ✗ none, SaaS only |
| EU pricing | ✗ list price × 1.15, plus $0.05/GB EU add-on | Same list price, EU site available |
| Best for | Few engineers, many hosts, high data volume | Many engineers, moderate host count, wide integrations |
Sources: New Relic pricing · New Relic usage plan · Datadog pricing. Annual-commitment rates; on-demand is higher on both. Compiled July 2026.
Where Datadog wins
Access is free, so the whole team can look
Datadog does not meter readers of infrastructure monitoring or APM. That single fact is why teams that hit New Relic's five-user Standard cap end up on a Datadog trial. Ten engineers who all need full New Relic access cost $3,490 a month in seats at the Pro list price of $349, before a single gigabyte. On Datadog those ten people cost nothing, and the same ten engineers running twenty hosts pay $300 for Infrastructure Pro and $620 for APM, so $920 in total. Support engineers, product managers and on-call staff outside the platform team can all open a dashboard without a procurement conversation.
Logs are the cheapest to ingest
At $0.10 per ingested or scanned gigabyte, Datadog ingest is a quarter of New Relic's $0.40. If most of your telemetry is application logs you keep for compliance and search rarely, the Logging without Limits approach of ingesting everything and indexing selectively is genuinely cheaper than paying New Relic's rate to ingest all of it.
The integration catalogue is the widest in the category
Whatever database, queue, cloud service or proxy you run, there is almost certainly a maintained Datadog integration with a dashboard and a set of default monitors. New Relic's coverage is broad but Datadog's is broader, and on a heterogeneous estate that difference shows up as weeks of work you do not have to do.
Where New Relic still wins
New Relic does not care how much you run. Hosts, containers, serverless functions and Kubernetes nodes are not billing units, so a small team operating a large estate pays for people and gigabytes only. Take three engineers and two hundred hosts: on New Relic Standard that is $208 a month in seats plus data, while on Datadog the same estate costs $3,000 for Infrastructure Pro and $6,200 for APM, so $9,200 a month before logs. Container-dense and autoscaling workloads are exactly where per-host pricing hurts most, and it is the most common reason teams move in the other direction.
The second advantage is that New Relic's bill has fewer surprises inside it. Log search is included in the ingest price rather than billed separately as indexed events, and custom metrics are counted as ingested gigabytes rather than as their own metered series. Datadog's per-module structure is what produces invoices several times the estimate, especially when a high-cardinality tag turns one OpenTelemetry metric into thousands of billable series. New Relic's free tier is also in a different league: 100 GB a month and a full platform user, against 5 hosts with one day of metric retention.
Migrating New Relic to Datadog
The path most teams take:
1. Count hosts before anything else. Include autoscaled instances at peak, not average, and remember that APM is priced per host on top of infrastructure. If the number is large relative to your headcount, stop here; the migration will not save money.
2. Swap the collection layer. Install the Datadog Agent, or point your existing OpenTelemetry collector at Datadog's OTLP intake. Both platforms accept OTLP, so teams already on OpenTelemetry change an endpoint and a key rather than re-instrumenting.
3. Model metric cardinality. Before cutting over, list the tags on your metrics. Infrastructure Pro includes 100 custom metrics per host and bills beyond that, so a user ID or container ID tag is where budgets break. This is the step teams skip and then regret.
4. What does not move. NRQL dashboards, alert conditions and service levels are rebuilt in Datadog's own syntax, and historical telemetry stays in New Relic until retention expires. Neither platform can be self-hosted, so this migration swaps one vendor dependency for another rather than removing it. If that is the thing you want to fix, look at SigNoz or Grafana Cloud instead.
Common questions
FAQ: New Relic vs Datadog
Is Datadog cheaper than New Relic?
Count engineers, then count hosts. Ten engineers on New Relic Pro cost $3,490 a month in seats alone before any data, while twenty hosts on Datadog with Infrastructure Pro and APM cost $920 with no seat fee. Reverse the shape and it reverses too: three engineers on New Relic Standard cost $208 a month no matter how many containers they run, whereas two hundred Datadog hosts with APM cost $9,200. Whichever number is larger, people or hosts, decides which model is cheaper.
Does Datadog charge per user like New Relic?
No, not for the core observability products. Datadog meters infrastructure and APM per host, logs per gigabyte and per indexed event, and several products per session or per test, so adding a reader to a dashboard costs nothing. Some newer modules do have per-user components, but the products that replace New Relic APM and infrastructure monitoring do not. That is the whole appeal for a team that hit New Relic's five-user Standard cap and does not want to reprice everyone at $349.
What is the custom metrics problem when moving from New Relic to Datadog?
Datadog's Infrastructure Pro plan includes an allotment of 100 custom metrics per host and bills beyond it. Any metric with a high-cardinality tag such as user ID, container ID or request path multiplies into many billable time series, and OpenTelemetry metrics frequently land in that bucket. On New Relic the same data is counted as gigabytes ingested at $0.40, so a metric-heavy OTel pipeline that looks cheap on New Relic can be the largest line on a Datadog invoice. Model your real tag cardinality before you commit.