Skip to content
Calcrivo

Azure Cosmos DB RU Calculator

Convert reads, writes and queries into Cosmos DB request units and price provisioned, autoscale and multi-region throughput.

Inputs

reads/s
writes/s
KB

A 1 KB point read costs 1 RU; a 1 KB write costs about 5 RU.

queries/s
RU

Read the actual charge from the request-charge header.

%
regions

Doubles the per-RU rate in every region.

USD/hour
GB
USD/GB-month

Monthly Cost

$580.52

Required Throughput

3,900RU/s

Billed Throughput

3,900RU/s

Throughput Charge

$455.52

Storage Charge

$125.00

RU per Write

10.0RU

Step by step

  1. Values used

    Point reads per second = 800 reads/s; Writes per second = 150 writes/s; Average item size = 2 KB; Queries per second = 60 queries/s; RU per query = 12 RU; Throughput mode = Standard provisioned; Average autoscale utilisation = 35 %; Regions = 2 regions; Multi-region writes enabled = No; Price per 100 RU/s = 0.0080 USD/hour; Transactional storage = 250 GB; Storage price per GB = 0.2500 USD/GB-month

  2. Azure Cosmos DB RU

    RU/s = reads × (1 RU per KB) + writes × (5 RU per KB) + queries × RU per query, rounded up to the next 100; cost = RU/s ÷ 100 × rate × 730 × regions, ×1.5 on autoscale and ×2 with multi-region writes.

  3. Monthly Cost

    = 580.52

  4. Required Throughput

    = 3,900 RU/s

  5. Billed Throughput

    = 3,900 RU/s

  6. Throughput Charge

    = 455.52

  7. Storage Charge

    = 125.00

  8. RU per Write

    = 10.0 RU

How it works

Cosmos DB normalises every operation into request units: a 1 KB point read is 1 RU, a 1 KB write about 5 RU because of index maintenance, and queries are whatever the request-charge header reports. Autoscale bills 1.5× the standard rate but only for the RU/s actually used between 10% and 100% of your maximum, so it wins when the peak-to-average ratio is worse than about 1.5. Provisioned throughput is billed per region every hour whether you use it or not, so an over-provisioned container replicated to three regions with multi-region writes costs six times what the numbers suggest — verify the per-100-RU/s rate for your region and API in the Azure pricing calculator.

Formula

Azure Cosmos DB RU

RU/s = reads × (1 RU per KB) + writes × (5 RU per KB) + queries × RU per query, rounded up to the next 100; cost = RU/s ÷ 100 × rate × 730 × regions, ×1.5 on autoscale and ×2 with multi-region writes.

RU
Request unit — the normalised cost of one database operation
autoscale floor
Autoscale never bills below 10% of the maximum RU/s you set
region factor
Provisioned throughput is billed in full in every region you replicate to

Frequently Asked Questions

How is Azure Cosmos DB RU calculated?

RU/s = reads × (1 RU per KB) + writes × (5 RU per KB) + queries × RU per query, rounded up to the next 100; cost = RU/s ÷ 100 × rate × 730 × regions, ×1.5 on autoscale and ×2 with multi-region writes. Cosmos DB normalises every operation into request units: a 1 KB point read is 1 RU, a 1 KB write about 5 RU because of index maintenance, and queries are whatever the request-charge header reports. Autoscale bills 1.5× the standard rate but only for the RU/s actually used between 10% and 100% of your maximum, so it wins when the peak-to-average ratio is worse than about 1.5.

Why does Azure Cosmos DB RU matter?

Provisioned throughput is billed per region every hour whether you use it or not, so an over-provisioned container replicated to three regions with multi-region writes costs six times what the numbers suggest — verify the per-100-RU/s rate for your region and API in the Azure pricing calculator.

What values do I need to enter?

This calculator takes 12 inputs: Point reads per second, Writes per second, Average item size, Queries per second, RU per query, Throughput mode, Average autoscale utilisation, Regions, Multi-region writes enabled, Price per 100 RU/s, Transactional storage, Storage price per GB. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.

You might also need