Datadog Annual Contracts vs On-Demand - How to Read the Three Unit Price Columns

Datadog's price list shows three unit prices side by side: annual contract, monthly contract, and on-demand. Following the official documentation, this article organizes what the three columns mean, the reality that how the gap is set changes with the type of SKU, how usage beyond the contracted volume is treated, and the order in which to build an estimate. For current unit prices, refer to the official pricing page.

Conclusion - The Three Columns Represent "Length of the Promise" and "Outside the Promise"

A single SKU on Datadog's price list shows three prices because, even for the same usage, the unit price changes with "how far ahead you promise." The annual contract is the unit price when you promise 12 months of usage, the monthly contract is the unit price when you promise volume on a month-by-month basis, and on-demand is the unit price applied to what you did not promise, that is, usage beyond the contracted volume and allotments.

What the three unit price columns mean and how they correspond to the official documentation (as of September 2026)
ColumnWhat it applies toCorrespondence in the official documentation
Annual contractVolume promised as 12 months' worthThe "contract price" displayed on the contract details screen
Monthly contractVolume promised on a monthly basisAlso the contract price. Higher than the annual contract because the promise period is shorter
On-demandUsage that fits within neither the contracted volume nor the allotmentsThe "on-demand price" on the contract details screen. On the usage screen it is distinguished by a purple on-demand display (beta)
Source: Billing / Usage Details (as of September 2026)

If you read the three columns of the price list as "cheap / normal / expensive," you will misjudge the contract form. Correctly, it is a two-tier mechanism in which "up to the promised volume is the contract price, and the overflow is the on-demand price," and designing to reduce the overflow is the theme of this article. The stacking structure of SKUs itself is left to How to Decode the Pricing Structure. This article explains how billing works. For current unit prices and plans, refer to Datadog's official pricing page.

How the Gap Is Set Changes with the Type of SKU

The gap between the three columns is not uniform across all SKUs. Comparing the three columns SKU by SKU on the official pricing page, the way the gap is set splits into three by the type of billing unit.

How the three unit price columns line up for major SKUs (by type)
SKU typeSKUs that belong to it (billing unit)Position of the monthly contractPosition of on-demand
Host-basedInfrastructure Monitoring Pro / Enterprise (host), APM (host), Database Monitoring (DB host)Higher than the annual contractSame as the monthly contract
Usage-proportionalLogs Standard Indexing (1 million events), Indexed Spans (1 million spans), Synthetics API / browser tests (runs), RUM Investigate / Product Analytics (1,000 sessions)Higher than the annual contractHigher still than the monthly contract (the type where the three columns spread the most)
All three columns equalLogs Ingest (GB)Same as the annual contractSame as the annual contract
Source for the pattern: the three columns of each SKU on Datadog Pricing. For specific unit prices, refer to the same page

Three patterns can be read from the table. For host-based SKUs (Infrastructure Monitoring, APM, Database Monitoring), the monthly contract and on-demand are the same price, and the gap from the annual contract is the smallest of the three types. For SKUs proportional to usage, such as log indexing, span indexing, Synthetics, and RUM, the monthly contract is higher than the annual contract and on-demand sits further on top of that, making this the type where the three columns spread the most. There are also SKUs like Logs Ingest where all three columns are the same price, and for those there is no point in agonizing over the contract form. Since the size of the gap itself can be revised, check the three columns with your own eyes on the official pricing page before signing.

Usage Beyond the Contracted Volume Shows Up in Purple Mid-Month

Usage beyond the contracted volume is distinguished as "on-demand" on the usage screen (as of September 2026, this display is a beta feature, and the official documentation states that using it requires a request to your account representative). According to the official documentation, the usage trend graph carries a blue display indicating the contracted and allotted portions and a purple display indicating the on-demand portion, and a dashed Committed line indicates the contracted volume for each product (as of September 2026; source: Usage Details). If you have requested the display, overages are not something you learn about only when the invoice arrives; they are visible mid-month as the purple part of the graph. A notification for the first use of a product not included in your contract (likewise a beta feature requiring a request) is also provided, and once enabled, an email arrives about 48 hours after usage begins. It is worth enabling right after signing the contract as a means of learning early about an unintended SKU lighting up. The specifics of how to read the usage screen and the order in which to check the cost items prone to spiking are left to How to Check Your Costs.

Host-Based Products Have Two Billing Plans

Host-based products such as Infrastructure Monitoring, APM, and Database Monitoring have two kinds of billing plans (as of September 2026; source: Pricing (Billing)). The High watermark plan (HWMP) records the host count every hour and determines the monthly charge from the maximum after excluding the top 1%, and is described as being for environments where usage is stable throughout the month. The Hybrid monthly/hourly plan (MHP) promises a monthly minimum volume and then bills host hours beyond it at an hourly rate, and is for environments with sharp increases and decreases such as autoscaling or short-lived workloads. It can be read as a contract in the form of "pay for the excess by the hour rather than by the month." The billing cycle begins on the 1st of every month (UTC) regardless of the contract date, and the first month is prorated (source: Billing). How hosts are counted is left to that entry; here we note only the point that "the more the host count moves, the more important it becomes to place the annual contract volume at the floor."

Allotments Are Another Layer Outside the Contracted Volume

Between the contracted volume and on-demand there is a layer called allotments. This is why the contracted portion and the allotment from the parent product are grouped together in the same blue display on the usage screen: APM hosts come with monthly allotments of Ingested Spans and Indexed Spans, and Infrastructure Monitoring hosts come with allotments of containers and custom metrics, in proportion to the host count (as of September 2026; source: Usage Details). The on-demand unit price applies only to the portion exceeding both the contracted volume and the allotments. Because of this layer, estimating usage-proportional SKUs is not a binary choice between contract and on-demand but a procedure of first counting "how much the parent product's allotment can absorb." The rule that APM allotments are pooled across all hosts and a worked example are in How to Think About APM Pricing, and how to choose between log indexing and the Flex Logs tier is in The Three Tiers of Log Pricing. If you decide the contracted volume without counting allotments, you end up either with leftover or with overflow.

Build the Estimate in Three Stages, from Free Measurement to an Annual Contract

Extracting only how to choose the contract form, the order comes down to the following three stages.

  • 1. Measure: Run a representative configuration on the Free plan or the free trial and obtain actual numbers for host count, ingested GB, and session count (the scope is in Free Plan and Trial)
  • 2. Monthly contract: Run 1 to 2 months on a monthly contract, recording each month's maximum and minimum to learn the range of variation
  • 3. Annual contract: Put only the "floor" volume you use every month without fail on an annual contract, and leave the peaks on monthly contract or on-demand. Review the floor at renewal
The three stages to deciding the contract form (unit price column source: Datadog Pricing)

The numbers to obtain in the measurement stage are limited to those that will later become the units of contracted volume. Capturing four of them, host count, log GB / day, monthly Indexed Spans, and session count, covers the majority of SKUs in the type table. The reason you might want to skip the second stage is that "a monthly contract is more expensive than an annual one," but recalling the type table changes the judgment. For host-based SKUs, the monthly contract and on-demand are the same price, so there is no "penalty for not promising" during the period of watching on a monthly contract. For usage-proportional SKUs, the monthly contract is cheaper than on-demand, so even during the measurement period it works out cheaper to promise volume on a monthly contract. Since the annual contract is a discount premised on continuing to use the promised volume for 12 months, the two key points of this order are not to sign one before measuring, and when you do sign, to set it at the floor volume rather than the peak.

SKUs Suited to Annual Contracts, and SKUs That Are Not

The simplification that "an annual contract makes everything uniformly cheaper" applies only to host-based SKUs, so we separate suited and unsuited SKUs from both the gap pattern and the allotments.

Suitability for annual contracts by SKU type
SKU typeGap patternConditions suited to an annual contractCaveats
Host-based (Infrastructure Monitoring / APM / Database Monitoring)Monthly = on-demand. Only the annual contract is lowerThe host count floor has been stable for several monthsPut only the floor on the annual contract. The fluctuating portion costs the same on on-demand as on a monthly contract
Usage-proportional (log indexing / Indexed Spans / Synthetics / RUM / Product Analytics)Three steps in which the unit price rises in the order annual, monthly, on-demandThe range of variation is understood from 2 months on a monthly contractThe type where overflow is most expensive. Set the contracted volume slightly above the floor, and first subtract what the allotments can absorb
All three columns equal (Logs Ingest)No differenceNo point in choosing a contract formThe only thing that can move the cost is controlling ingestion volume
Unit price column source: Datadog Pricing. Source for the explanation of allotments: Usage Details

The way to use this is to see which type the top three cost items on your invoice fall into. Thinking in the host-based type, let the annual-contract host unit price be p and the on-demand unit price be q (q > p): if you have 20 Pro hosts in a busy month and 15 in a quiet one, holding 15 hosts on the annual contract and 5 on on-demand costs 15 × p + 5 × q per month. That is cheaper than the 20 × q of holding all 20 on a monthly contract (for host-based SKUs, monthly = on-demand), the difference from the 20 × p of putting all 20 on the annual contract is 5 × (q − p), and in exchange, in months when you do not use the 5 hosts you do not pay for them. If host-based SKUs are at the top, the annual contract decision boils down to determining the host count floor. If usage-proportional SKUs are at the top, narrowing the volume that spills into on-demand through allotments and ingestion controls before an annual contract has more effect on the invoice than the choice of contract form. The three prices in the columns are not "options arranged from cheapest," but parts you combine to fit the shape of your own usage.

References

Share