New Relic bills on three meters at once: platform users, data ingest, and synthetic checks. The free edition covers one full user, 100 GB of ingest a month and 500 synthetic checks. Past that, ingest is $0.40 per GB, full platform users are $10 for the first and $99 for each additional on Standard (maximum five) or $349 each on Pro with an annual commitment, and synthetic checks above your edition allowance are $0.005 each.
That is the short version, and it is already more moving parts than most vendors have in total. The reason the question "how much does New Relic cost" gets asked so often is that there is no single answer on the pricing page. There is a rate card, and your bill is whatever your usage multiplies it into.
So here is the arithmetic, meter by meter, with the numbers you actually need to model a bill before you sign anything.
The three meters, and which one usually bites
New Relic restructured years ago away from per-host pricing toward a usage model, and the shape it landed on is unusual: you are charged for the people who look at the data and for the volume of data itself, rather than for the number of servers you point it at. For a company with a lot of hosts and a small platform team, that is a genuinely good deal. For a large engineering organisation where everyone wants access, the seat meter is the one that hurts.
| Meter | Included | Rate beyond the included amount |
|---|---|---|
| Data ingest | 100 GB per month, every edition | $0.40 per GB, or $0.60 per GB on the Data Plus option |
| Full platform users | 1 on the free edition | $10 first user, then $99 each up to 5 on Standard; $349 per user on Pro with annual commitment, $418.80 monthly pay as you go |
| Core users | None included beyond the free user | $49 per user on every edition |
| Synthetic checks | 500 Free, 10,000 Standard, 1M Pro, 10M Enterprise | $0.005 per check |
| Simple ping monitors | Unlimited on every edition | No charge, they sit outside the check meter |
Note the gap between Standard and Pro on the user line. Standard caps you at five full platform users at $99 each. The moment you need a sixth, you are not paying $99 for one more seat, you are moving to Pro at $349 per user for all of them. That is the single largest cliff in New Relic's pricing and it catches teams that grow past a handful of engineers.
What the free edition actually includes
New Relic's free tier is a real product rather than a trial, and it is more generous than most of the market. One full platform user, 100 GB of data ingest every month, 500 synthetic checks, and unlimited simple ping monitors, with no expiry and no card required. For a solo developer or a small side business, that can genuinely be the whole bill.
The catch is the shape of the limits rather than their size. One full user means one person can use the parts of the platform that matter; everyone else needs a $49 core seat or a $99 full seat. And 100 GB sounds like a lot until you turn on log forwarding, at which point a busy application can eat it in days.
How much does New Relic synthetic monitoring cost?
This is the meter people model worst, because the headline number looks trivial. Checks above your allowance are $0.005 each. Half a cent. The problem is the multiplier.
An average month is 30.44 days, or 43,833 minutes. New Relic's fastest monitor frequency is once per minute, so a single monitor running flat out from one location is 43,833 checks a month. At $0.005 each, that is about $219 for one monitor, if you are paying the overage rate. And checks are counted per location: New Relic's own documentation works the example of a monitor running once an hour from five locations, which comes to 24 times 5, or 120 checks a day.
| One monitor, this schedule | Checks per month | Cost at $0.005 per check |
|---|---|---|
| Every minute, 1 location | 43,833 | $219.16 |
| Every minute, 3 locations | 131,499 | $657.50 |
| Every 5 minutes, 1 location | 8,767 | $43.84 |
| Every 5 minutes, 3 locations | 26,301 | $131.50 |
| Every 15 minutes, 3 locations | 8,766 | $43.83 |
Now put the allowances back in, because they change the answer completely. Pro includes 1 million checks a month, which covers 22 monitors running every minute from a single location, or 114 monitors on a five-minute interval, before a cent of overage. Enterprise includes 10 million, or 228 such monitors. Standard's 10,000 checks, on the other hand, is consumed in under four hours by a single one-minute monitor.
So the practical rule is: on Pro and Enterprise, price synthetics on the allowance and largely forget the rate. On Standard, price them on the rate, because you will be paying it all month.
The ping monitor exemption is the most valuable thing here
Buried in the documentation is a line worth more than most of the pricing page. The check limits "apply to all monitor types except the simple ping monitor", and ping monitors are included in every New Relic edition at no extra cost. Unlimited. On the free edition included.
If your requirement is plain uptime monitoring, send a request to a URL and tell me when it stops answering, New Relic will do that for nothing, at a one-minute interval, from multiple locations, forever. There is no other observability platform at this size offering that.
The boundary matters though. A ping monitor is a request and a response code. The moment you want to assert that a keyword appears in the body, check a JSON field, verify a certificate, or walk a login flow, you are on a non-ping monitor type and the meter starts. Those results also generate telemetry that counts against your 100 GB, so a large fleet of checks has an ingest cost even when the checks themselves are free. If the ingest line on your bill is the one climbing, the useful first move is working out which pipelines are actually producing that volume before you start trimming retention.
Three worked examples
Rate cards are abstract, so here are three realistic shapes with the arithmetic shown. All figures are US dollar list prices checked in August 2026, and all assume you stay inside 100 GB of ingest unless stated.
A solo developer with a side project
One person, six services to keep an eye on, no log forwarding. This fits inside the free edition outright: one full user at $0, ingest well under 100 GB, and six simple ping monitors at $0 because pings are unmetered. Total: $0 a month. This is the case where New Relic is unbeatable and you should just take it.
A five-engineer startup on Standard
Five full platform users: $10 for the first, $99 for the other four, so $406 a month. Say 250 GB of ingest with logs turned on, which is 150 GB over at $0.40, adding $60. Twelve ping monitors at $0, plus four scripted API checks running every five minutes from two locations, which is 4 times 2 times 8,767, or 70,136 checks, against a 10,000 allowance, so 60,136 over at $0.005, adding about $301. Total: roughly $767 a month, and note that the synthetic checks are nearly 40% of it despite being the smallest part of the product they use.
The same team after moving to Pro
Five users at $349 on an annual commitment is $1,745 a month, which is a large step up. But the Pro allowance of 1 million checks swallows those 70,136 checks whole, so the synthetics line drops to $0, and the ingest overage stays at $60. Total: roughly $1,805 a month. More than double, for headroom on users and checks they may not need yet. This is exactly the decision point where teams start pulling uptime monitoring out into a cheaper dedicated tool and keeping New Relic for APM.
Where New Relic gets expensive, and what to do about it
Three patterns account for most surprise invoices. The first is log forwarding switched on without a retention policy, which turns a 100 GB allowance into a four-figure ingest line. The second is the Standard-to-Pro user cliff at six people. The third is scripted synthetic checks left at a one-minute interval across several locations, where the multiplier does the damage quietly.
The fixes are unglamorous. Sample or drop noisy log sources at the agent rather than after ingest. Audit who genuinely needs a full platform seat against a $49 core seat. And be deliberate about check intervals: dropping a scripted check from one minute to five cuts its cost by 80% and, on most endpoints, costs you four minutes of detection time you were probably not acting on anyway.
That last point deserves an honest caveat. If you genuinely need fast detection, lengthening the interval is the wrong lever, because you are trading away the thing you bought monitoring for. The better move is to separate the two jobs: keep deep instrumentation where you need traces, and run plain availability checks somewhere they are not metered per execution.
New Relic vs Datadog on cost
These two land on the same shortlist constantly, and the cost comparison is not one-directional. On the marginal metered check, Datadog is much cheaper: about $0.0005 per API test run against New Relic's $0.005, a factor of ten. On plain uptime pings, New Relic is cheaper by an infinite margin, because they are free and Datadog meters every execution.
Which one wins therefore depends entirely on what your checks look like. Heavy scripted API assertions at high frequency favour Datadog. A wide fleet of simple availability pings favours New Relic. The full side-by-side, including the included allowances and the interval floor both share, is on our Datadog vs New Relic comparison, and the Datadog side is broken down in what Datadog synthetics cost.
Is New Relic worth it for uptime monitoring alone?
If you are already on New Relic and you want basic availability checks, yes, obviously: the ping monitors are free and you are not adding a vendor. It is hard to argue with free.
If uptime is the only thing you want and you are considering adopting New Relic for it, the answer is usually no, and for reasons that have nothing to do with price. The interval floor is one minute, so an outage can run 60 seconds before the first failed check. Alerting is something you assemble from conditions and workflows rather than something that arrives configured, and SMS or voice needs an integration rather than being in the box. The interface is built for an engineer debugging a distributed system, which is a poor fit for a marketing lead who wants to know whether the site is up.
That is the trade dedicated tools exist to make. We check every 30 seconds on the Team plan and above, confirm an outage from three regions before paging anyone, and include unmetered SMS, email, Slack and webhook alerts on every plan for a flat $19 to $189 a month with no ingest meter and no per-run counter. We do not do APM, tracing, logs or real user monitoring, so if you need those, New Relic is not something we replace. Plenty of teams run both, deliberately.
Frequently asked questions
Is New Relic free?
Yes, genuinely. The free edition gives one full platform user, 100 GB of data ingest a month, 500 synthetic checks and unlimited simple ping monitors, with no time limit. Additional full users are $10 for the first and $99 each after on Standard, and ingest past 100 GB is $0.40 per GB.
How much does New Relic cost per user?
A core user is $49 on every edition. A full platform user is $10 for the first and $99 for each additional on Standard, capped at five users. On Pro it is $349 per user with an annual commitment, or $418.80 per user on monthly pay as you go. Enterprise pricing is quoted per customer.
How much does New Relic charge for data?
Every edition includes 100 GB of data ingest a month. Beyond that it is $0.40 per GB on the standard data option, or $0.60 per GB on Data Plus, which adds longer retention and additional compliance features. Synthetic check results, logs, traces and metrics all count toward the same pool.
Does New Relic do 30-second checks?
No. New Relic's synthetic monitor frequencies start at every minute, then step out to 5, 10, 15 and 30 minutes and longer. There is no sub-minute option to buy on any edition. Datadog holds the same one-minute floor, so if you need faster detection than that, both platforms are out.
What is the cheapest way to run New Relic?
Stay on the free edition if one full user covers you, use $49 core seats instead of $99 full seats wherever someone only needs to read dashboards, keep availability checks on simple ping monitors so they stay unmetered, and control log volume at the agent rather than paying to ingest it and then dropping it.