Your Datadog renewal just landed on your desk and the number made you do a double take. You are not alone. Datadog pricing has become the single biggest budget conversation in most engineering orgs we talk to, and the story is almost always the same: the bill grew 30 to 50% year over year, nobody can fully explain why, and finance wants answers by Friday.
This post breaks down the Datadog pricing model in plain language, shows you where the costs actually hide, and walks through what mid-market teams are doing about it.
How Datadog Pricing Works: A Full Breakdown by SKU
Datadog does not sell one product. It sells around twenty, each with its own meter. You are not buying observability; you are buying a bundle of SKUs that each charge on a different unit. Here is what the core ones cost:
Datadog bills infrastructure monitoring per host: Pro is $15 (annual) and Enterprise is $23. Datadog bills based on the 99th percentile host count, so misconfiguring the agent to run in every Kubernetes pod can multiply your host count overnight.
APM is also per host. Standard is $31, Pro is $35, and Enterprise is $40 per host per month (annual). Each tier covers around 1 million indexed spans per host; Datadog meters anything beyond that separately.
Log management is where most teams get surprised. Ingestion costs $0.10 per GB, and indexing is charged separately based on retention: $1.06 per million events for 3 days up to $2.50 for 30 days.
Custom metrics are billed at $0.05 per metric per month. High-cardinality tags on a single Prometheus-style metric can explode into thousands of distinct custom metrics fast.
On top of these, add-ons like RUM, Database Monitoring, CI Visibility, and Session Replay each carry their own meter. That is where the bill starts compounding.
Why Is My Datadog Bill So High?
The sticker price is not really the problem. The problem is how Datadog cost behaves over time. Three things compound:
Cardinality growth.Every new microservice, tag or deployment environment multiplies the number of unique time series. A single high‑cardinality tag such as user_id or request_id can balloon custom metric counts by orders of magnitude.
Log volume drift. Engineers leave debug logs on in production. New services ship with verbose defaults. Teams set retention policies once and never revisit them. Because Datadog meters ingestion and indexing separately, even a few months of extra retention can add thousands of dollars to the bill.
SKU sprawl. Somebody enabled RUM for a pilot. Somebody else turned on Database Monitoring. CI Visibility got added during an incident retro. None of these got switched off, and each one has its own growth curve. Because Datadog bills are based on the 99th‑percentile host count and each add‑on has its own meter, temporary spikes or misconfigurations, like running the agent in every pod rather than on each node, can multiply your host count by tenfold.
The net effect is a 30 to 50% year over year increase that arrives right around renewal, with enough pricing complexity that nobody on your team can cleanly defend it to the CFO.
Is Datadog Worth the Cost in 2026?
Sometimes yes. For small teams who value a single pane of glass and do not want to run anything themselves, Datadog is genuinely a good product. Where it stops making sense is when the bill grows faster than your infrastructure budget. Analysts note that mid‑size teams with 60 hosts and 50 to 150 GB/day of logs spend around $20,000 to $25,000 per month on Datadog. When monthly observability exceeds roughly 5% of your infrastructure cost or pushes past the $20,000 mark, about $240,000 annually, the economics tilt toward running an open‑source stack. At that scale, teams switching to a managed or self‑hosted open‑source stack often see savings in the 50 to 75% range.
Datadog Alternatives: Open Source Stack That Cuts Costs by 75%
Prometheus for metrics, Loki for logs, Grafana for dashboards, and OpenTelemetry for instrumentation. This stack covers everything Datadog does, without the vendor lock-in.
A self-hosted setup for 50 hosts runs $400 to 800 per month. Managed open-source typically runs $2,000 to 4,000 per month. Either way, that is 50 to 75% less than a typical Datadog bill.
The honest tradeoffs: you need someone who knows how to operate it, the initial migration takes real engineering effort, and you lose some of the polish of Datadog’s newer features. For most mid market teams spending $15K a month or more, the math still works out heavily in favor of migrating.
Here is how the monthly cost actually compares at three common deployment sizes:
| Deployment Size | Datadog (all SKUs, typical) | Open Source Self-Hosted | Open Source Managed | Typical Savings |
| Small (~20 hosts) | ~$4,856/mo | $300 to 500/mo | $1,200–2,000/mo | 60–90% |
| Mid-size (~60 hosts, 50–150 GB/day logs) | ~$20,988/mo | $500–900/mo | $2,500–4,500/mo | 75–88% |
| Large (~150 hosts) | ~$42,468/mo | $1,200–2,000/mo | $5,000–9,000/mo | 75–88% |
How to Migrate from Datadog to Open Source (Step by Step)
A typical migration from Datadog to an open‑source stack runs about 8–12 weeks for a mid‑market environment, and usually requires a team of two to three engineers committing around 200 to 400 hours. The sequence we follow:
- First, we audit your current Datadog usage and map out which SKUs are driving cost and which are actually being used.
This alone usually reveals 15 to 25% of spend that can be cut before any migration even starts.
- Next, we instrument with OpenTelemetry so your applications emit telemetry in a vendor neutral format.
This decouples you from Datadog permanently.
- Finally, we stand up the Prometheus, Loki, and Grafana stack, migrate dashboards and alerts with fidelity, and run both systems in parallel until the team trusts the new one.
Finally, we turn Datadog off and your next renewal is a fraction of what it used to be.
Want to know what you'd save on open source?
We'll model your workload and give you a real number
Book a 30 minute callFrequently Asked Questions
How much does Datadog cost on average?
Datadog costs range from roughly $4,856 per month for a small deployment to over $42,000 per month for larger environments. A mid-size team with 60 hosts and 50 to 150 GB/day of logs typically spends around $20,988 per month. Cost per host across all SKUs ranges from $40 to $80 in large clusters to $160 or more in smaller, log-heavy deployments.
Are there good alternatives to Datadog?
Yes, open-source stack (Prometheus, Loki, Grafana, OpenTelemetry) together offer a proven alternative that costs 50 to 75% less than Datadog. A self-hosted version for around 50 hosts costs $400 to 800 per month, and a managed service typically runs $2,000 to 4,000 per month. Skedler specializes in moving teams from Datadog to open source and has helped companies cut observability costs by up to 70%.
How long does a Datadog migration take?
With Skedler managing the process, a mid-market Datadog migration typically takes 8 to 12 weeks. It usually requires two to three engineers committing around 200 to 400 hours. Skedler compresses that timeline significantly by handling the instrumentation, dashboard migration, and parallel running phases for you rather than leaving them to be built from scratch.
When does it make sense to move off Datadog?
Skedler recommends evaluating a move once monthly observability spend exceeds $20,000 per month or roughly 5% of your infrastructure budget. At that threshold, an open-source stack can cut costs by 50 to 75% and pay for itself within a year. Skedler offers a free cost audit to show you exactly what you would save before you commit to anything.
What is the cheapest Datadog plan?
The cheapest paid Datadog plan is Infrastructure Pro at $15 per host per month billed annually. The Free tier covers up to five hosts but is only practical for evaluation. Once you add APM, logs, and any add-ons, the effective cost per host climbs considerably above the base rate.
When does it make sense to move off Datadog?
The economics usually tip in favor of migrating once monthly observability spend exceeds around 5% of your infrastructure budget or $20k per month (about $240k annually). In that range, an open‑source stack can cut costs by 50–75% and pay for itself within a year.
How can i reduce my Datadog bill?
Skedler helps teams migrate from Datadog to open source and has helped companies cut observability costs by up to 70%. The process starts with a free cost audit that maps exactly which Datadog SKUs are driving your bill and what the open-source equivalent would cost at your scale.
