Datadog's 2025 Pricing Restructure: The Honest Numbers
Datadog's mid-2025 pricing changes were sold as 'simplification.' We modeled the restructured SKUs against three real customer environments. The simplification is real for two of them. The third paid 31% more for the same telemetry.
In this review
| Criterion | Score |
|---|---|
| Editorial Score | 3.7 |
| Value for Money | 3.4 |
| Implementation Effort | 4.5 |
| Vendor Trajectory | 4.0 |
| Overall | 3.90 / 5.00 |
↑ What works
- +The restructured SKU bundles do simplify procurement for most mid-market customers
- +The integrated platform remains genuinely best-in-class for the modal customer
- +Custom metrics and trace ingestion got noticeably more transparent in the new pricing
↓ Where it disappoints
- −Customers with heavy log volumes are paying meaningfully more under the new model
- −The 'unlimited' bundle qualifiers have caveats that require careful contract review
- −Negotiating power has shifted toward the vendor; renewals are visibly tighter
Observability vendor pricing is a genre. Datadog has been the leading practitioner for nearly a decade and the mid-2025 pricing restructure is, depending on which slide you read, either a customer-friendly simplification or the latest move in the long arc of how a usage-based-pricing vendor extracts more revenue per customer per year. Reasonable people can disagree on the framing. We modeled the new pricing against three real customer environments to answer the only question that matters: did your bill go up or down?
What changed
The new SKU structure consolidates from a granular per-feature line-item model to a smaller number of bundle SKUs (Pro, Pro Plus, Enterprise, Enterprise Plus) with tier-specific feature sets. Custom metrics, trace ingestion, and several previously-add-on observability surfaces are now bundled into the higher tiers. Log management remains usage-priced, with tiered indexing and retention models.
The headline pitch is "simpler procurement, predictable spend." For most customers, that pitch is approximately accurate. For customers with heavy log volumes, the pitch obscures a real price increase.
The three test environments
We modeled three customers — anonymized but real — against the new pricing.
The first: a 180-engineer SaaS company with moderate log volumes (8 TB/month indexed), 320 tracked services, and roughly 12 million custom metrics per month. Old-pricing annual cost: $612,000. New-pricing annual cost: $584,000. Net savings: 4.6%. The customer benefits from the bundle math because they consume nearly all the included features.
The second: a 95-engineer fintech with light log volumes (2 TB/month) and a heavy traces footprint. Old: $284,000. New: $268,000. Net: −5.6%. Same dynamic.
The third: a 320-engineer logistics platform with heavy log volumes (44 TB/month indexed at 30-day retention), moderate traces, and aggressive custom-metric usage. Old: $1.4M. New: $1.83M. Net: +30.7%.
The pattern is that the bundle math rewards the customer whose usage profile fits the bundle Datadog defined. The customers whose usage skews heavily into log management — which remained granularly-priced and which had several thresholds tightened — pay materially more.
The bundle math rewards customers who fit the shape Datadog optimized for. The customers who don't fit the shape pay more.
The qualifier issue
The most consequential change at the contract level is the addition of fair-use qualifiers on bundles previously presented as "unlimited." Specifically, custom metrics in the Pro and Pro Plus tiers now have a fair-use threshold that triggers at volumes a real engineering organization will hit at scale. Above the threshold, custom metrics revert to per-metric pricing.
In our reading of the new contract templates, this is a structural change worth taking seriously. The threshold is not always disclosed up-front in sales conversations and the contractual language is in the SKU appendix rather than the main agreement. Buyers should ask for the threshold to be quoted explicitly and, ideally, raised in their negotiation.
On the alternatives
Grafana Cloud, Honeycomb, and the ELK-stack-self-hosted option are all real alternatives. None is a drop-in replacement for the integrated experience Datadog provides. The migration cost from Datadog to any of them is meaningful — typically 4 to 8 engineering-months, depending on the depth of the existing Datadog footprint — and the ongoing operational overhead of running self-hosted observability stacks is real. The migration math has improved in the last 18 months but has not yet crossed the threshold where most existing Datadog customers should make the move.
The exception is the very-heavy-log-volume customer (10 TB/month and up) who is paying a step-function more under the new pricing. For that customer, the migration math now produces a defensible business case for the move.
The verdict
Datadog at 2025 pricing remains the right answer for most observability buyers, particularly those whose usage profile matches the bundle definitions. For heavy-log customers and customers with idiosyncratic usage shapes, the procurement conversation has hardened materially. Renewals will require more rigorous modeling than they have in past cycles. The vendor has earned its position; the customer should still negotiate as if the position were contestable.
- Devin J.
Our log volume is the issue. We pay 28% more for what's effectively the same telemetry. Considering Grafana Cloud for the second time.
- Marie L.
The bundle qualifiers are the part that bites. 'Unlimited custom metrics' has a fair-use clause that triggers at volumes a real org will hit.
- Naomi B. (author)
@Marie — yes, that fair-use clause is the buried lede of the restructuring. Buyers should ask for it in writing before signature.
- Theo F.
Has anyone moved off Datadog successfully? We've talked about it for 18 months and the migration always looks too painful.
- Aviva R.
The 'integrated platform' point is real. We tried to go best-of-breed and ended up paying more for less coherent data.
The Weekly Briefing
Did this review help?
Get one of these on your desk every Monday morning. Free, opinionated, no sponsored items.
