The hard part of choosing among Datadog alternatives is the billing unit. Datadog charges per host and per million log events. SigNoz charges per GB. Honeycomb sells event bundles, groundcover charges per host in your own cloud, and New Relic charges per GB plus per user. The same workload lands on a different bill under each. The ranking below is our judgment, for a team of 10-50 engineers replacing Datadog, and every figure comes from the vendor's own pricing page, read on 28 October 2026. These are list prices with no discounts. Observability is one of many stack decisions; the rest are in Hire developers by tech stack: rates, vetting and interview guides.
Datadog alternatives at a glance
| Platform | Billing unit | Published starting price | Free option | Best for |
|---|---|---|---|---|
| SigNoz | Per GB of logs and traces, per million metric samples; no host or user pricing | Logs and traces $0.3/GB; metrics $0.1 per million samples; Teams from $49/month | No free cloud tier on the page; self-hosted open source; startup offer $19/month for 12 months | The closest like-for-like swap, with an open source exit |
| Grafana Cloud | Per series, per GB, per active user, plus a $19/month platform fee | Metrics $6.50 per 1k series; logs and traces $0.05 process, $0.40 write and $0.10 retain per GB; $8 per active user | 10k series, 50 GB logs, 50 GB traces, 3 users | Teams already on Prometheus or OpenTelemetry |
| New Relic | Per GB ingested, plus per user | $0.40/GB beyond 100 GB/month; Core user $49; Full platform Pro user $349 (annual) | 100 GB/month and one free full-platform user | Few seats, lots of data |
| Honeycomb | Event volume bundles | Pro from $150/month (up to 750M events) | 20M events/month | Debugging distributed traces |
| groundcover | Per host, running in your cloud | Pro $30/host/month; Enterprise $35 | Free plan with 12-hour retention | Kubernetes fleets that must keep data in their own VPC |
Best for is our judgment; prices are from each vendor's pricing page in October 2026.
SigNoz is first because it's the most direct replacement for Datadog's three signals under one roof, and because you can walk away from the hosted product by self-hosting the open source version. That's judgment, not a benchmark. Datadog isn't a row because it's the baseline.
Why is Datadog so expensive?
Start with what Datadog's own pricing page says, because most Datadog alternatives pages don't quote it. Infrastructure Pro starts at $15 per host per month billed annually, or $18 on-demand. Enterprise is $23 annually or $27 on-demand. APM adds $31 per host per month when attached to infrastructure monitoring and billed annually, with APM Pro at $35 and APM Enterprise at $40. Standalone APM is $36, per Datadog's FAQ.
Logs are billed in two pieces. Ingest starts at $0.10 per ingested or scanned GB per month. Standard indexing with 15-day retention is $1.70 per million log events per month on annual billing, or $2.55 on-demand.
The bill grows through the units you don't watch. Datadog's FAQ says Infrastructure Pro includes 100 custom metrics per host (200 on Enterprise), and that custom metrics are "counted hourly and averaged monthly across your entire account, not on a per-host basis." So one team's metric habit raises everyone's bill.
Two things the page doesn't say, at least in the portion we read: a price for monthly commits, and anything about how hosts are counted at peak. You'll see claims about Datadog's billing mechanics elsewhere. We aren't repeating any we couldn't find on Datadog's page. Treat any figure for Datadog's "typical annual spend" the same way unless it names its source.
What 20 hosts and 100 GB of logs cost on each price list
This is an illustration built only from the unit prices quoted above and below, not a quote. It leaves out discounts, custom metrics, retention beyond defaults and egress. Your bill will differ.
Hosts first. Datadog Infrastructure Pro on annual billing: 20 x $15 = $300 a month. On-demand: 20 x $18 = $360. Add APM at $31 per host: 20 x ($15 + $31) = $920 a month annual. groundcover Pro: 20 x $30 = $600 a month. groundcover runs in your cloud, so the storage in your account is extra, and the groundcover pricing page we read doesn't give a number for it.
Logs at 100 GB a month. Datadog's ingest fee is 100 x $0.10 = $10. But indexing is billed per million events, and on the pages we read, Datadog's price list gives no conversion from GB to events, so the full log bill can't be computed from the price list. That mismatch is the point. A team that logs small, chatty lines pays differently from one that logs fat JSON, even at the same GB. SigNoz: 100 x $0.3 = $30. New Relic: 100 GB falls inside its free 100 GB a month, so ingest is $0, with users extra.
Seats. New Relic Core users are $49 each, so 5 x $49 = $245 a month. Grafana Cloud charges $8 per active user, so 10 users with no free seats is 10 x $8 = $80. Grafana's free plan includes 3 active users and the page implies they're included, so 7 x $8 = $56 is likely. $80 is the conservative figure.
Per-host and per-GB pricing cross over somewhere that depends on your log volume per host, which is why Datadog's list price can look fair for 20 hosts and painful once logs grow. Per-GB meters are spreading, too. See Cloudflare's per-GB log and trace pricing from 1 December for another vendor moving to a volume meter.
Datadog vs New Relic and Grafana: which bills lower?
Neither page lets you declare a winner without your own numbers. Datadog bills hosts and log events; New Relic bills ingested GB beyond 100 GB a month at $0.40, plus users. Standard Full platform seats are $10 for the first user and $99 for each additional user up to 5, and Pro is $349 per user annual ($418.80 monthly). If you have five engineers and a great deal of data, New Relic's seat prices are what you'll feel.
Grafana Cloud has a free plan and a $19 platform fee to start, but it has four meters (series, log GB, trace GB, users), and each needs a forecast. Comparing it with Datadog means counting active series, which you can pull from Prometheus today.
Is there a free alternative to Datadog?
Yes, for small workloads. Datadog's own Free tier covers up to five Basic hosts with 1-day metric retention, which is a trial more than a tier.
Among the alternatives: Grafana Cloud's free plan includes 10k active series, 50 GB of logs, 50 GB of traces and 3 active users. New Relic includes 100 GB a month of ingest and one free full-platform user. Honeycomb's free tier is 20M events a month. groundcover has a free plan, with 12-hour retention, which isn't enough for investigating yesterday's incident. SigNoz's pricing page lists no free cloud tier, but the open source version is free to self-host, and the startup offer is $19 a month for 12 months.
The catch with self-hosting is that the software is free and the operator isn't. Someone has to upgrade it, size the storage and get paged when it falls over. If that's your plan, read setting up an on-call rotation for a nearshore team first, because your monitoring now needs monitoring.
What is better than Datadog?
It depends on the situation. This table is judgment, built from the prices above.
| Your situation | Pick | Why |
|---|---|---|
| You want one product with the same three signals | SigNoz | Per-GB pricing with no host or user meter |
| You already run Prometheus or OpenTelemetry | Grafana Cloud | Free tier covers 10k series, and you know the data |
| Few engineers, large data volumes | New Relic | First 100 GB a month is free, seats stay small |
| Your pain is debugging traces | Honeycomb | Event-based, with a 20M-event free tier |
| Kubernetes, and data can't leave your VPC | groundcover | Per-host price, runs in your cloud |
| Your bill is small and the team is happy | Stay on Datadog | Five free Basic hosts, and moving costs engineer time |
Staying is a legitimate answer. In our judgment, if your 20-host bill is $920 a month, switching saves little compared with an engineer's week.
What this means for your observability budget
Decide in this order. First, find out what your Datadog bill is made of. Pull the last three invoices and split them into hosts, log events and custom metrics. If custom metrics or indexed logs dominate, you have a fix inside Datadog before you need a new vendor: index fewer logs, drop unused metrics, and ask for a quote on committed use.
Second, pick the alternative whose unit matches your growth. If you're adding hosts, a per-GB vendor helps. If you're adding log volume, a per-host vendor does. That's judgment, and the 20-host illustration above is the template for testing it with your own numbers.
Third, price the people. A self-hosted stack saves licence cost and spends engineer time, and a platform move is a migration project with an owner, dashboards to rebuild and alerts to re-tune. For that side of the decision, see the labor behind a platform migration.
Hiring DevOps engineers for an observability migration
If nobody on your team wants to own the move, bring someone in who has done it. HighCircl matches vetted engineers in 72 hours, from seven European countries, at €45-105/hr ($50-115/hr) with a 20% capped margin and no subscription or recruitment fee. About 1 in 10 applicants passes vetting. Ask for DevOps engineers through HighCircl with observability migration experience.
FAQ
Why is Datadog so expensive?
It bills several units at once. Datadog's pricing page lists $15 to $23 per host per month for infrastructure, $31 to $40 for APM on top, $1.70 per million indexed log events, and $0.10 per GB ingested. Custom metrics are counted across your whole account, not per host. The cost comes from stacking those units, not from any one price.
Is there a free alternative to Datadog?
Yes, for small workloads. Grafana Cloud, New Relic and Honeycomb have free tiers, and SigNoz is open source if you host it yourself. Datadog's own free tier covers five Basic hosts with 1-day metric retention. In our judgment, none of the free tiers replace a paid plan for production on a team of 10-50.
What is better than Datadog?
It depends on your situation. In our judgment SigNoz is the closest like-for-like replacement, Grafana Cloud fits Prometheus shops, and New Relic suits few seats with lots of data. See the table above.
Datadog vs New Relic: which is cheaper?
The units differ, so it depends on your workload. Datadog bills per host and per million log events. New Relic bills per GB beyond 100 GB a month at $0.40, plus users at $49 for Core and $349 for Pro Full platform seats on annual billing. There's no per-user price on the Datadog pages we read, so the seat count is what moves the comparison.
Datadog vs Grafana: which is cheaper?
Grafana Cloud, not the self-hosted version, has a free plan and a $19 platform fee to start, with metrics at $6.50 per 1k series and $8 per active user. Beyond that it depends on your series count and log volume, which you'd need to measure.