Azure Cosmos DB vs Firestore
Compare Cosmos DB and Firestore for document-backed applications, SDK and API scope, partitioning, free allowances and consumption costs.
Which database approach fits the workload?
Consider Cosmos DB when its selected NoSQL account, capacity model and Azure integration fit the application. Consider Firestore when its selected edition, interface and application SDK workflow fit the project. Compare the exact APIs rather than assuming all document services behave alike. Client access, real-time updates, recovery and free quota all depend on the retained configuration.
Seven differences that change the shortlist
| Buying question | Azure Cosmos DB | Google Cloud Firestore |
|---|---|---|
| Selected service | This brief focuses on Cosmos DB’s NoSQL document service | Specify Firestore Standard or Enterprise and database mode |
| Application interface | Confirm the intended API, SDK and account | Confirm the intended edition’s API, SDK and client capabilities |
| Data design | Partition-key and query design remain application work | Document, collection, index and query design remain application work |
| Capacity | Provisioned, autoscale and serverless arrangements differ | Edition-specific operation or unit charging plus storage and transfer |
| Access control | Define application identity and effective database permissions | Define client or server identity and effective access controls |
| Free boundary | Eligible provisioned account; free tier excludes serverless | Limited eligible free quota; certain recovery features require billing |
| Regional scope | Specify regions, consistency and recovery requirements | Specify location, edition and recovery requirements |
Four worked buying briefs
These are hypothetical scenarios with explicit assumptions. They do not establish equal capabilities, measured usability or a negotiated price.
| Requirement | Proposed exercise | Accept when |
|---|---|---|
| Data boundary | Attempt cross-account record and list access | Unauthorised documents remain inaccessible |
| Update workflow | Edit a record across two allowed sessions | Observed behaviour meets the written application brief |
| Query scope | Add the same filtered application screen | Indexes, partition assumptions and usage are understood |
| Recovery | Exercise a permitted sandbox recovery scenario | Required data and app access can be restored under the retained offer |
A private record
Can one user read only their allowed documents?
Azure Cosmos DB: Implement the actual application and database identity path.
Google Cloud Firestore: Implement the actual client or server identity path.
Test a copied document identifier and a list query with two isolated accounts.
An application update
How does an edited record reach another authorised session?
Azure Cosmos DB: Exercise the proposed SDK and update-delivery implementation.
Google Cloud Firestore: Exercise the selected edition and SDK update workflow.
Specify the actual behaviour instead of inferring real-time or offline support from the database name.
A new query
What work and usage does a new filtered screen require?
Azure Cosmos DB: Inspect the partition and index approach and request-unit demand.
Google Cloud Firestore: Inspect required indexes and the selected edition’s operation charging.
Keep the same authorised result and query frequency.
A recoverable deletion
Can a sandbox record or database be recovered as required?
Azure Cosmos DB: Use the purchased account and recovery configuration.
Google Cloud Firestore: Use the purchased backup or recovery feature scope.
Free access to ordinary operations does not prove recovery is included.
Choose the API before choosing the quota
Firestore now has Standard and Enterprise editions with different query and compatibility scope. Cosmos DB account and API choices also matter. Record the exact interface used by the application and the features it depends on.
For direct client access, inspect authentication and effective record permissions. For server access, inspect the server’s credentials and authorisation logic. A successful administrator query does not validate the user boundary.
Free allowances cover different things
Cosmos DB’s documented free tier applies to an eligible provisioned account and is not available for serverless. Firestore’s free quota has database and usage conditions, while backup, restore and point-in-time recovery require billing.
Model the actual read, write, query, listener or update-delivery workload, storage and transfer in the selected editions. Include regional and recovery requirements before comparing a small free prototype with production.
Questions before you choose
Does Cosmos DB free tier apply to serverless accounts?
No. Microsoft documents the free tier for eligible provisioned accounts; serverless is excluded.
Are Firestore Standard and Enterprise interchangeable?
Do not assume that. The editions have different query, compatibility and billing scope; validate the actual SDK and API requirements.
Does a free database quota include every backup and restore?
No. Check the exact service. Firestore specifically lists backup, restore and other recovery features as requiring billing.
Your practical evaluation pack
Download the database evaluation checklist (CSV)
Four proposed checks with blank evidence and owner fields. Record the exact version, edition, region, deployment and workload used.
Turn the comparison into your own evaluation
Use the same criteria and priorities for both products, keep unknowns unassessed and record the plan and evidence behind your ratings.
Build a Azure Cosmos DB vs Google Cloud Firestore scorecard →Keep the assumptions with the figures
Use the date, billing cycle, units and limits with each example. Check current account terms before purchasing. Unverified prices and model costs stay unpriced.
Download the buying examples as JSON ↓Share the comparison
Use these examples when discussing the shortlist with your team.