Observability / APM

New Relic

New Relic, Inc.

Single-platform, consumption-priced observability that's cheap for big fleets/small teams but charges per data-GB and per full-user seat.

Category
Observability / APM
License
Proprietary
Deployment
SaaS
Cost
Medium
Free tier
Yes
Self-host effort
Maturity
Incumbent
Popularity
Long-standing top-3 APM incumbent

The catch

The per-user "full platform" seat ($349/user) is the trap — a sharp cliff from Standard to Pro, NRQL creates lock-in, and the ingest model still drives aggressive sampling to control cost.

Monitors

ServersMetricsLogsTracesSyntheticsCloudK8s

Capabilities

AlertingDashboardsDistributed tracingAnomaly detectionAPI

Built for

Mid-marketEnterpriseDevelopers

The honest take

New Relic is the observability incumbent most worth re-examining if your mental model of it is a decade old. In 2020 it threw out per-host pricing — the thing everyone hated about it and still hates about Datadog — and rebuilt around two meters: data ingested (priced per GB) and full-platform users (per seat). That single change is why it keeps showing up as the cheaper answer for large or elastic fleets: adding hosts is free, so an autoscaling estate that would set Datadog’s high-water-mark meter on fire barely moves the New Relic bill. The free tier is genuinely generous too — 100 GB/month and one full user, enough to actually run a small production system on, not a toy.

The catch is the seat model, and it’s a sharp one. The “full platform” user that can do everything lists around $349/month, and the jump from a Standard/basic user to a full Pro user is a cliff, not a ramp. For a big fleet watched by a handful of engineers, that math is fantastic. For a 60-person engineering org where everyone wants to poke at dashboards during an incident, the per-user line can become the dominant cost and the thing you ration access over — which is exactly the wrong incentive for observability. Before you switch, count your users, not your hosts.

The other quiet cost is ingest discipline. Because you pay per GB in, the platform nudges you toward sampling and dropping data to control spend — which works, but means you’re making “what do we keep?” decisions that a flat-rate tool wouldn’t force. And NRQL, New Relic’s query language, is powerful and genuinely pleasant, but it’s also lock-in: dashboards and alerts written in it don’t port anywhere, so the more you invest the more switching costs you build.

Net: if you’re leaving Datadog specifically because the per-host bill is unpredictable, New Relic is one of the most credible SaaS landing spots — same “one platform, full SaaS, no servers to run” convenience, very different cost curve. Just model the seat count honestly, because that’s where the surprise lives. The head-to-head with Datadog lays out exactly which fleet shapes favour which bill.

First-hand data

data as of Jun 24, 2026

Significant incidents · 90d
31 major · 2 minor
Incident-minutes logged
15 hcumulative, not downtime
Last incident
May 18, 2026

Polled first-hand from each vendor's public status page & GitHub. "Significant" excludes informational notices & planned maintenance; incident-minutes sum per-incident durations (not platform downtime). Method & full data →