AWS
Microsoft Azure
Google Cloud
Key takeaways
1
Provisioned throughput: Pay hourly for allocated RU/s, even when idle.
2
Autoscale: Pay for the highest throughput reached each hour.
3
Serverless: Pay for consumed RUs and storage, without pre-provisioning throughput.
4
Additional costs: Storage, replication, backups, and data transfer can increase bills.
Azure Cosmos DB Pricing at a Glance
For RU-based Azure Cosmos DB APIs, the main throughput billing models are standard provisioned throughput, autoscale provisioned throughput, and serverless.The right option depends on how frequently your application accesses data, how predictable demand is, and whether you need multi-region deployment.
For a broader breakdown of Azure compute, databases, storage, and networking costs, see our Azure pricing guide.
| Pricing model | Billing basis | Best suited for |
|---|---|---|
| Standard provisioned | Allocated RU/s per hour | Consistent, predictable workloads |
| Autoscale provisioned | Highest scaled RU/s each hour | Variable workloads with traffic spikes |
| Serverless | RUs consumed | Intermittent or low-volume applications |
| Free tier | Eligible throughput and storage allowance | Development and small applications |
| vCore-based offerings | Provisioned compute and storage | Supported MongoDB and other vCore architectures |
For the RU-based pricing models, Microsoft provides the following published or documented pricing examples.
| Cost component | Example rate or allowance |
|---|---|
| Standard provisioned throughput | $0.008 per 100 RU/s-hour |
| Autoscale, single-write region | 1.5× the standard throughput rate |
| Serverless requests | $0.25 per 1 million RUs |
| Transactional storage | $0.25 per GB-month |
| Free tier | First 1,000 RU/s and 25 GB, subject to eligibility |
Pricing assumptions: The $0.008 per 100 RU/s-hour standard provisioned rate is an East US example from Microsoft’s Cosmos DB billing documentation. The 1.5× autoscale multiplier assumes a single-write-region account. The $0.25 per million serverless RUs and $0.25 per GB-month storage figures are illustrative amounts in Microsoft’s cost-planning guide, not verified current regional quotes. The throughput examples assume RU-based APIs, standard provisioned or autoscale capacity as indicated, and no availability-zone surcharge. Prices exclude discounts and additional charges unless stated otherwise. Confirm the chosen region, account settings, and current billing meters in the Azure Pricing Calculator.
The official Azure Cosmos DB pricing page provides the applicable pricing models and regional configurations.
Unlike a conventional database where compute may be priced by virtual machine size, the RU-based Cosmos DB APIs use Request Units to measure the resources required to perform database operations.
For example, reading a 1 KB item typically costs 1 RU, while more complex queries can consume significantly more.
See Microsoft’s Request Units documentation for details.
The official Azure Cosmos DB pricing page provides the applicable pricing models and regional configurations.
How Azure Cosmos DB Pricing Works
Azure Cosmos DB pricing is primarily determined by database operations and storage consumption.Unlike a conventional database where compute may be priced by virtual machine size, the RU-based Cosmos DB APIs use Request Units to measure the resources required to perform database operations.
What Are Request Units (RUs)?
A Request Unit (RU) measures the processing resources consumed by Cosmos DB operations, including reads, writes, updates, and queries. RU consumption depends on document size, indexing, and query complexity.For example, reading a 1 KB item typically costs 1 RU, while more complex queries can consume significantly more.
See Microsoft’s Request Units documentation for details.
| Database operation | Illustrative RU cost | Microsoft's planning assumptions |
|---|---|---|
| Point read | 1 RU | 1 KB item, read by ID and partition key |
| Create item | 5 RUs | Average for a 1 KB item with fewer than 5 indexed properties |
| Update item | 10 RUs | Average for a 1 KB item with fewer than 5 indexed properties |
| Delete item | 5 RUs | Illustrative planning estimate; actual indexing affects cost |
| Indexed query | 10 RUs | Query uses indexing effectively and returns 100 results or fewer |
These are Microsoft’s cost-planning estimates, not fixed charges. Actual RUs depend on item size, indexing, consistency, and query complexity. Measure real request charges for production workloads.
Provisioned throughput charges for allocated RU/s, while serverless charges for consumed RUs.
Azure bills for the allocated capacity on an hourly basis, even when your application is idle.
What Is RU/s in Azure Cosmos DB?
Request Units per second (RU/s) measure available database throughput. For example, 1,000 RU/s can support approximately 1,000 one-RU operations per second under suitable conditions.Provisioned throughput charges for allocated RU/s, while serverless charges for consumed RUs.
Azure Cosmos DB Pricing Models Compared
Azure Cosmos DB provides different billing models for steady workloads, unpredictable traffic, and applications with infrequent database activity.1. Standard Provisioned Throughput Pricing
Standard provisioned throughput lets you allocate a fixed amount of RU/s to a database or container.Azure bills for the allocated capacity on an hourly basis, even when your application is idle.
Monthly throughput cost = (Provisioned RU/s ÷ 100) × Hourly rate per 100 RU/s × Hours
For example, using Microsoft’s illustrative $0.008 rate per 100 RU/s-hour:
| Provisioned throughput | Estimated monthly cost (730 hours) |
|---|---|
| 400 RU/s | $23.36 |
| 1,000 RU/s | $58.40 |
| 5,000 RU/s | $292.00 |
| 10,000 RU/s | $584.00 |
These estimates assume one Azure region and exclude storage and networking. Standard provisioned throughput suits predictable workloads, but allocated RU/s is billed even when unused.
You configure a maximum throughput (Tmax), and autoscale adjusts between 10% and 100% of that maximum. For a 4,000 RU/s maximum, the minimum is 400 RU/s. Billing is based on the highest throughput reached during each hour, subject to the configured minimum.
For a typical single-write-region account, Microsoft documents an autoscale pricing multiplier of 1.5 times the standard provisioned throughput rate.
2. Azure Cosmos DB Autoscale Pricing
Autoscale automatically adjusts provisioned throughput according to demand.You configure a maximum throughput (Tmax), and autoscale adjusts between 10% and 100% of that maximum. For a 4,000 RU/s maximum, the minimum is 400 RU/s. Billing is based on the highest throughput reached during each hour, subject to the configured minimum.
For a typical single-write-region account, Microsoft documents an autoscale pricing multiplier of 1.5 times the standard provisioned throughput rate.
Autoscale hourly cost = (Highest RU/s reached ÷ 100) × Standard hourly rate × 1.5
Consider a container configured with a maximum of 4,000 RU/s.
If it scales to 1,000 RU/s during an hour, the illustrative charge is:
(1,000 ÷ 100) × $0.008 × 1.5 = $0.12
If it reaches 4,000 RU/s during another hour:
(4,000 ÷ 100) × $0.008 × 1.5 = $0.48
At the 400 RU/s minimum, the illustrative charge is (400 ÷ 100) × $0.008 × 1.5 = $0.048/hour, even with no requests. A brief peak can set the charge for the whole hour. See Microsoft’s autoscale FAQ.
Autoscale charges based on the highest throughput reached each hour. It suits variable workloads, but sustained high utilization may make standard provisioned throughput more economical.
Microsoft’s cost-planning documentation illustrates a serverless request rate of $0.25 per million RUs.
If it scales to 1,000 RU/s during an hour, the illustrative charge is:
(1,000 ÷ 100) × $0.008 × 1.5 = $0.12
If it reaches 4,000 RU/s during another hour:
(4,000 ÷ 100) × $0.008 × 1.5 = $0.48
At the 400 RU/s minimum, the illustrative charge is (400 ÷ 100) × $0.008 × 1.5 = $0.048/hour, even with no requests. A brief peak can set the charge for the whole hour. See Microsoft’s autoscale FAQ.
Autoscale charges based on the highest throughput reached each hour. It suits variable workloads, but sustained high utilization may make standard provisioned throughput more economical.
Dynamic Autoscale vs. Traditional Autoscale
Azure Cosmos DB supports dynamic autoscale, which independently adjusts throughput across regions and physical partitions. This can reduce costs when traffic varies between regions or partitions. Traditional autoscale scales throughput more uniformly, potentially increasing costs when demand is uneven.3. Azure Cosmos DB Serverless Pricing
Serverless charges for the Request Units consumed by database operations instead of continuously allocated throughput.Microsoft’s cost-planning documentation illustrates a serverless request rate of $0.25 per million RUs.
Serverless monthly cost = (Monthly RUs ÷ 1,000,000 × RU rate) + Storage charges
For example:
| Monthly request consumption | Request charges |
|---|---|
| 1 million RUs | $0.25 |
| 10 million RUs | $2.50 |
| 20 million RUs | $5.00 |
| 100 million RUs | $25.00 |
| 500 million RUs | $125.00 |
These estimates exclude storage and additional services.
Serverless can be attractive for applications with long idle periods, development environments, and workloads with low average request volumes.
However, it has important limitations.
Serverless can be attractive for applications with long idle periods, development environments, and workloads with low average request volumes.
However, it has important limitations.
Serverless accounts support a single Azure region rather than global multi-region replication.
Serverless does not qualify for the standard Cosmos DB free-tier throughput allowance.
Throughput and performance limits differ from provisioned throughput.
Sustained high-volume workloads may cost less under provisioned throughput.
Review Microsoft’s provisioned throughput versus serverless comparison before choosing a capacity model.
For related analytics workloads, see our Azure Databricks pricing guide.
The following examples show how these components affect an estimated monthly bill.
4. vCore-Based Azure Cosmos DB Pricing
Azure Cosmos DB for MongoDB vCore uses compute and storage-based pricing rather than RU-based billing. Costs depend on provisioned resources and cluster configuration.For related analytics workloads, see our Azure Databricks pricing guide.
Azure Cosmos DB Pricing Per Month: Three Examples
Monthly Azure Cosmos DB costs depend on the selected pricing model, the workload’s throughput requirements, and the amount of stored data.The following examples show how these components affect an estimated monthly bill.
Example 1: Small Application Using Serverless
Consider a small, single-region serverless application with occasional database activity, using Microsoft’s illustrative request and storage rates rather than a live regional quote.| Pricing input | Example value |
|---|---|
| Monthly request consumption | 20 million RUs |
| Serverless RU price | $0.25 per million RUs |
| Data and index storage | 10 GB |
| Storage price | $0.25 per GB-month |
Estimated monthly cost:
Requests: 20 million RUs × $0.25/million = $5.00
Storage: 10 GB × $0.25/GB = $2.50
Total: $7.50/month
This illustrative estimate excludes additional networking, backup, and other applicable charges.
Assume East US standard provisioned throughput, 50 GB of transactional storage, 730 hours, and no availability-zone surcharge or discounts. The storage price is illustrative.
Example 2: Production Application Using Provisioned Throughput
Consider a production application requiring 1,000 RU/s continuously.Assume East US standard provisioned throughput, 50 GB of transactional storage, 730 hours, and no availability-zone surcharge or discounts. The storage price is illustrative.
| Pricing input | Example value |
|---|---|
| Provisioned throughput | 1,000 RU/s |
| Hourly price per 100 RU/s | $0.008 |
| Monthly runtime | 730 hours |
| Storage | 50 GB |
| Storage rate | $0.25/GB-month |
Estimated monthly cost:
Throughput: (1,000 ÷ 100) × $0.008 × 730 hours = $58.40
Storage: 50 GB × $0.25/GB = $12.50
Total: $70.90/month
Provisioned throughput is billed even when the allocated RU/s is not fully used. Additional charges are excluded.
Assume autoscale is configured with a maximum of 4,000 RU/s per region and 100 GB of data and indexes replicated into both regions.
Assume a single-write account with a maximum of 4,000 RU/s and a minimum of 400 RU/s in each region. Both regions are assumed to follow the same hourly scaling pattern and the same illustrative $0.012 per 100 RU/s-hour rate (based on the East US standard example). This is a modeled scenario, not a live quote for two specific regions; it excludes availability-zone surcharges and discounts.
This example models traditional autoscale with identical throughput scaling in both regions. Dynamic autoscale may produce a different bill because each region and physical partition can scale independently. The estimate should not be interpreted as a universal multi-region autoscale price.
Example 3: Autoscale With Two Azure Regions
Consider a production application deployed across two Azure regions with a single write region.Assume autoscale is configured with a maximum of 4,000 RU/s per region and 100 GB of data and indexes replicated into both regions.
Assume a single-write account with a maximum of 4,000 RU/s and a minimum of 400 RU/s in each region. Both regions are assumed to follow the same hourly scaling pattern and the same illustrative $0.012 per 100 RU/s-hour rate (based on the East US standard example). This is a modeled scenario, not a live quote for two specific regions; it excludes availability-zone surcharges and discounts.
This example models traditional autoscale with identical throughput scaling in both regions. Dynamic autoscale may produce a different bill because each region and physical partition can scale independently. The estimate should not be interpreted as a universal multi-region autoscale price.
| Throughput reached per region | Hours/month |
|---|---|
| 400 RU/s | 500 |
| 1,000 RU/s | 180 |
| 4,000 RU/s | 50 |
| Total | 730 |
Estimated monthly cost:
Using the illustrative autoscale rate of $0.012 per 100 RU/s-hour:
Using the illustrative autoscale rate of $0.012 per 100 RU/s-hour:
Low traffic: (400 ÷ 100) × 500 × $0.012 = $24.00
Medium traffic: (1,000 ÷ 100) × 180 × $0.012 = $21.60
Peak traffic: (4,000 ÷ 100) × 50 × $0.012 = $24.00
Throughput across two regions: $69.60 × 2 = $139.20
Replicated storage: 100 GB × 2 × $0.25 = $50.00
Total: $189.20/month
This simplified example assumes identical scaling and rates across both regions. Actual charges vary by configuration and exclude additional backup, networking, and other applicable costs.
Storage, backup configuration, replication, and network traffic can introduce additional charges.
Azure Cosmos DB Storage and Additional Costs
Throughput is often the largest pricing consideration, but it is not the only component of an Azure Cosmos DB bill.Storage, backup configuration, replication, and network traffic can introduce additional charges.
| Cost component | How it affects your bill |
|---|---|
| Transactional storage | Billed for consumed data and indexes |
| Analytical storage | Additional storage charges where enabled |
| Backup storage | Additional charges can apply based on backup configuration |
| Multi-region replication | Throughput and storage billed across regions |
| Data transfer | Applicable Azure bandwidth and egress charges |
| Dedicated gateway | Separate charges when provisioned |
| Availability-zone configuration | Charges depend on the selected model and configuration |
Transactional Storage Pricing
Azure Cosmos DB charges for stored data and indexes. At Microsoft’s illustrative rate of $0.25 per GB-month, 100 GB costs approximately $25 monthly in one region.Storage replicated across multiple regions increases the total billed capacity. Actual rates depend on the selected region and configuration.
Multi-Region Replication Costs
Azure Cosmos DB charges for provisioned throughput and replicated storage across configured regions.For example, 100 GB replicated across three regions represents approximately 300 GB of billed storage, or $75/month at an illustrative $0.25/GB-month.
Additional throughput and applicable bandwidth charges can increase the total.
Backup and Network Charges
Azure Cosmos DB supports backup and recovery capabilities, but additional backup storage or restore-related charges can apply depending on the selected configuration.Similarly, data leaving Azure or crossing supported network boundaries can incur bandwidth charges.
Use the official Azure Cosmos DB pricing page and Azure bandwidth pricing to estimate these components.
Azure Cosmos DB Free Tier: What Is Included?
Azure Cosmos DB provides an eligible lifetime free tier for supported provisioned-throughput accounts.The free tier includes:
The first 1,000 RU/s of provisioned throughput.
The first 25 GB of storage.
Availability for one eligible free-tier account per Azure subscription.
Support for eligible standard and autoscale provisioned-throughput configurations.
The allowance applies to the account rather than independently to every container.
Usage above the included limits is charged at applicable rates.
Usage above the included limits is charged at applicable rates.
Important:
Azure Cosmos DB serverless accounts are not eligible for the standard Cosmos DB free tier.
Free-tier accounts can still incur charges for services outside the allowance, including certain networking, backup, and additional configuration costs.
For development and small applications, the free tier can be an effective way to reduce initial database spending.
Review Microsoft’s Cosmos DB free-tier documentation for eligibility and billing examples.
The correct choice depends on traffic consistency, required performance, and how much capacity remains unused.
For development and small applications, the free tier can be an effective way to reduce initial database spending.
Review Microsoft’s Cosmos DB free-tier documentation for eligibility and billing examples.
Standard vs. Autoscale vs. Serverless: Which Is Cheaper?
No single Cosmos DB pricing model is cheapest for every application.The correct choice depends on traffic consistency, required performance, and how much capacity remains unused.
| Consideration | Standard | Autoscale | Serverless |
|---|---|---|---|
| Main billing basis | Allocated RU/s | Highest hourly scaled RU/s | Consumed RUs |
| Idle-period charges | Yes | Minimum throughput charge | No request charge without operations |
| Best traffic pattern | Steady | Variable or spiky | Intermittent |
| Multi-region support | Yes | Yes | No |
| Capacity configuration | Manual | Automatic within limits | No provisioned RU/s |
| Main cost risk | Unused capacity | Expensive peak hours | High request consumption |
Standard provisioned throughput generally suits steady workloads, autoscale handles variable demand, and serverless can be economical for intermittent applications.
Compare total monthly costs alongside performance requirements, regional availability, and throughput limits before choosing a model.
Compare total monthly costs alongside performance requirements, regional availability, and throughput limits before choosing a model.
How to Calculate Azure Cosmos DB Costs
The most reliable way to estimate Azure Cosmos DB pricing is to use actual workload requirements and Microsoft’s pricing tools.Azure Cosmos DB Monthly Cost Formula
For standard provisioned throughput:
Monthly cost = Throughput charges + Storage charges + Additional charges
Where:
Throughput charges = (RU/s ÷ 100) × Hourly throughput price × Billable hours × Applicable regional/configuration factors
For serverless:
Monthly cost = (Consumed RUs ÷ 1,000,000 × Rate per million RUs) + Storage + Additional charges
For autoscale, calculate throughput charges using the highest billed RU/s for each hour rather than multiplying a monthly average by a single rate.
Use the Azure Cosmos DB Pricing Calculator
Use the Azure Pricing Calculator to estimate monthly Cosmos DB costs:
1
Select Azure Cosmos DB and the applicable API and pricing model.
2
Choose your deployment region and account configuration.
3
Enter provisioned RU/s, autoscale maximum throughput, or estimated serverless RU consumption.
4
Add transactional storage, replicated regions, and applicable backup or networking charges.
5
Compare the estimate with your workload's expected usage and any eligible commitment discounts.
For more accurate capacity estimates, use Microsoft’s Cosmos DB cost-planning documentation. Existing deployments should also be checked in Azure Cost Management.
Why Azure Cosmos DB Bills Can Be Higher Than Expected
Cosmos DB bills can exceed initial estimates when throughput, storage, or replication requirements differ from the original assumptions.| Cost driver | Why it increases costs |
|---|---|
| Overprovisioned RU/s | Charges continue even when throughput is unused |
| Expensive queries | Complex operations consume more RUs |
| Inefficient partitioning | Hot partitions can create performance bottlenecks |
| Excessive indexing | Additional write RUs and storage |
| Autoscale spikes | Higher hourly peaks increase billed throughput |
| Multiple regions | Additional throughput and replicated storage |
| Growing data volume | Higher storage and potentially throughput requirements |
Reducing RU consumption improves efficiency, but provisioned-throughput costs may remain unchanged until allocated capacity is reduced.
For broader strategies, see our Azure cost management best practices.
Steady workloads may benefit from standard provisioned throughput, while variable workloads may be better suited to autoscale.
For infrequent database activity, evaluate serverless where its performance and deployment limits are acceptable.
Use point reads when possible, review expensive queries, and remove unnecessary indexed paths where appropriate.
Reducing RU consumption can help avoid throttling and lower the capacity required to serve the workload.
This can create hot partitions and limit effective throughput utilization.
Review partition-key distribution and query access patterns before increasing total RU/s.
If an application consistently uses less throughput than allocated, evaluate whether the configured RU/s can be reduced safely.
For autoscale, review the maximum throughput setting and hourly scaling behavior.
Reservations are available with one-year and three-year terms.
Microsoft documents discounts of 20% for certain one-year reservation configurations and 30% for certain three-year configurations, with higher savings possible at qualifying reservation sizes.
However, the final discount depends on reservation size, pricing configuration, and eligibility.
Reserved capacity is not a universal discount for every Cosmos DB charge.
It does not automatically cover serverless request consumption, transactional storage, or networking charges.
Before purchasing, compare expected eligible throughput usage with the reservation commitment and review the applicable terms.
Azure offers two distinct commitment options for eligible database spending. They are not interchangeable with the Azure Savings Plan for Compute.
For broader strategies, see our Azure cost management best practices.
How to Reduce Azure Cosmos DB Costs
Cost optimization should begin with database efficiency and capacity planning before purchasing long-term commitments.1. Choose the Right Throughput Model
Compare the actual workload pattern against the billing model.Steady workloads may benefit from standard provisioned throughput, while variable workloads may be better suited to autoscale.
For infrequent database activity, evaluate serverless where its performance and deployment limits are acceptable.
2. Optimize Queries and Indexing
Database operations consume RUs according to the resources they require.Use point reads when possible, review expensive queries, and remove unnecessary indexed paths where appropriate.
Reducing RU consumption can help avoid throttling and lower the capacity required to serve the workload.
3. Review Partition Keys and Data Distribution
Poor partition-key selection can concentrate traffic on a small portion of the database.This can create hot partitions and limit effective throughput utilization.
Review partition-key distribution and query access patterns before increasing total RU/s.
4. Reduce Unused Provisioned Throughput
Monitor actual throughput consumption against provisioned capacity.If an application consistently uses less throughput than allocated, evaluate whether the configured RU/s can be reduced safely.
For autoscale, review the maximum throughput setting and hourly scaling behavior.
5. Evaluate Azure Cosmos DB Reserved Capacity
Microsoft offers reserved capacity for eligible provisioned-throughput Cosmos DB workloads.Reservations are available with one-year and three-year terms.
Microsoft documents discounts of 20% for certain one-year reservation configurations and 30% for certain three-year configurations, with higher savings possible at qualifying reservation sizes.
However, the final discount depends on reservation size, pricing configuration, and eligibility.
Reserved capacity is not a universal discount for every Cosmos DB charge.
It does not automatically cover serverless request consumption, transactional storage, or networking charges.
Before purchasing, compare expected eligible throughput usage with the reservation commitment and review the applicable terms.
Azure Savings Plan for Databases vs. Cosmos DB Reserved Capacity
Azure offers two distinct commitment options for eligible database spending. They are not interchangeable with the Azure Savings Plan for Compute.
| Feature | Cosmos DB Reserved Capacity | Azure Savings Plan for Databases |
|---|---|---|
| Commitment | Specified provisioned throughput capacity | Fixed hourly spend on eligible database usage |
| Term | 1 or 3 years | 1 or 3 years |
| Coverage | Matching Cosmos DB provisioned throughput | Eligible billing meters across supported Azure databases, including Cosmos DB |
| Best for | Predictable Cosmos DB throughput | Eligible database spending that may shift across services |
| Limitation | Unused reserved throughput does not carry forward | Unused hourly spend does not carry forward; not every database charge qualifies |
Azure Savings Plan for Databases offers one-year and three-year spending commitments for eligible Azure database services, including Azure Cosmos DB. Discounts apply to qualifying usage within the hourly commitment, while additional usage is billed at applicable on-demand rates. Eligibility depends on the specific billing meter and agreement. Compare the savings plan with Cosmos DB reserved capacity before committing.
Confirm eligible usage and compare expected savings before purchasing. For compute commitments, see our Azure Savings Plans guide.
Review metrics such as normalized RU consumption, provisioned throughput, throttled requests, storage growth, and regional usage.
These metrics can help identify whether rising costs are caused by inefficient database operations, increasing demand, or infrastructure configuration.
For teams evaluating platforms to monitor and reduce broader Azure spending, see our comparison of the Best Azure cost optimization tools.
While Cosmos DB costs can be reduced through throughput, indexing, and query optimization, we focus on commitment-based savings across supported Azure services.
With CoPilot and Autopilot, teams can review or automate eligible commitment purchasing decisions without modifying workloads.
For more ways to reduce cloud spending, explore our Azure cost management best practices.
Confirm eligible usage and compare expected savings before purchasing. For compute commitments, see our Azure Savings Plans guide.
6. Monitor Costs and Set Budgets
Use Azure Cost Management to monitor Cosmos DB spending, identify unexpected increases, and configure budget alerts.Review metrics such as normalized RU consumption, provisioned throughput, throttled requests, storage growth, and regional usage.
These metrics can help identify whether rising costs are caused by inefficient database operations, increasing demand, or infrastructure configuration.
For teams evaluating platforms to monitor and reduce broader Azure spending, see our comparison of the Best Azure cost optimization tools.
How Usage.ai Fits Into Azure Cost Optimization
At Usage.ai, we help Azure teams identify and manage eligible cloud commitments to reduce infrastructure costs.While Cosmos DB costs can be reduced through throughput, indexing, and query optimization, we focus on commitment-based savings across supported Azure services.
With CoPilot and Autopilot, teams can review or automate eligible commitment purchasing decisions without modifying workloads.
For more ways to reduce cloud spending, explore our Azure cost management best practices.
Frequently asked questions
How much does Azure Cosmos DB cost per month?
Azure Cosmos DB monthly costs depend on throughput, storage, regions, and the pricing model. Using Microsoft's illustrative rates, a serverless application consuming 20 million RUs with 10 GB of storage costs approximately $7.50 per month. A single-region provisioned-throughput workload with 1,000 RU/s and 50 GB of storage costs approximately $70.90 per month, excluding additional charges.
How much does 1,000 RU/s cost in Azure Cosmos DB?
At an illustrative standard provisioned-throughput rate of $0.008 per 100 RU/s-hour, 1,000 RU/s costs $0.08 per hour or approximately $58.40 over 730 hours. Storage, additional regions, and other applicable services increase the total bill. Actual pricing depends on region and configuration.
Is Azure Cosmos DB free?
Azure Cosmos DB offers an eligible free tier that includes the first 1,000 RU/s of provisioned throughput and 25 GB of storage. The allowance applies to one eligible account per subscription. Additional usage and services may incur charges. Serverless accounts do not qualify for the standard Cosmos DB free tier.
Is Azure Cosmos DB serverless cheaper than provisioned throughput?
Serverless can be cheaper for intermittent workloads because it charges for consumed Request Units rather than continuously provisioned throughput. For sustained high-volume applications, provisioned throughput may be more economical. The right choice also depends on performance requirements and regional availability.
How can I reduce Azure Cosmos DB pricing?
Start by choosing the appropriate throughput model, optimizing queries and indexing, reviewing partition-key distribution, and reducing unused provisioned capacity. Then evaluate reserved capacity for predictable eligible throughput and use Azure Cost Management to monitor actual spending.
Share
Facebook
X
LinkedIn
Reddit
