Datadog API Cost Calculation
Updated over a week ago

Finout offers a structured breakdown of Datadog costs by Datadog products and additional tags.

Basic Cost API

Finout uses an estimated cost summary API for a detailed product cost breakdown. Each day, the API reports the total cost accumulated so far in the month. For example, on the 3rd of the month, the data reflects the total for the 1st, 2nd, and 3rd days combined. To isolate the cost for just the 3rd, we subtract the cumulative cost up to the 2nd from that of the 3rd. This difference gives us the specific cost for the 2nd day.

On-Demand vs. Commitment Costs

Datadog on-demand costs

These represent pay-as-you-go expenses, accruing and displaying on the relevant days when on-demand usage occurs.

Datadog commitment costs

Commitment costs are usually displayed on the 1st of every month, while on-demand costs will be presented on the day they occur.

Commitment costs can be categorized into different models:

  • Usage-based- In this model, your commitment costs are based on a predetermined “bank” of usage. Once you run out of your usage, on-demand charges kick in.

  • Host/ Hourly commitment- An hourly commitment is set, beyond which on-demand rates are applicable.

Finout’s Model for Commitment Costs Across the Month

API Cost (“Unblended Costs”)

Unblended costs represent the expenses as received from Datadog. In the case of commitments, these are shown on the 1st of the month, while on-demand costs are displayed on the relevant day of usage.

“Amortized” Costs

Finout uses the usage API to divide these costs across all days of the month to spread out the commitment cost throughout the month.

  • Usage-based- Finout’s algorithm calculates daily costs based on usage, adapting the rate based on historical and current monthly data. Since the discounted rate isn't reflected in the API and may vary, Finout's dynamic adjustment, leveraging the data received through the APIs, ensures the most current and precise cost representation.

  • Host- The commitment is evenly divided across the month's total hours, ensuring a consistent and understandable cost breakdown.

Reasoning for Infra host, APM hosts, and Timeseries cost spikes on the 1st of the month

  1. Infra host, APM host products- The monthly cost is updated daily based on the top 99th percentile of usage, with the final cost confirmed on the last day of the month.

    Datadog tracks hourly host counts for Infrastructure and APM billing, with end-of-month billing based on the 99th percentile usage. The billable host count is calculated using the highest count in the lower 99% of usage hours, excluding the top 1% to mitigate the impact of usage spikes.

  2. Timeseries (custom metrics)- The billable count for custom metrics is determined by averaging the number of custom metric hours over the month. Your monthly billable count for custom metrics (reflected on the Usage page) is calculated by taking the total of all distinct custom metrics recorded each hour throughout the month and dividing this month’s cumulative hours to receive a monthly average value.

Did this answer your question?