Atlas Data Federation GCS
Parent: MongoDB Atlas on GCP · Published reference · snapshot 2026-10-24 · skill mongodb-cohort-a7061a319c83
↓ Facts as markdownall context files
Source-anchored research for Atlas Data Federation GCS. Run status: COMPLETED.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
Measurements and reference values
- MongoDB documents separate Data Processed and Data Returned charges for GCS queries. Its example contrasts scanning a whole 10 GB file with reading a targeted 1 GB partition; savings depend on targeting the relevant partition. Query limits can restrict processed data per instance and project. (low confidence) (verified as of 2026-10-14) [source]
Problems, failure modes and limitations
- A federated database instance must run in Google Cloud to query GCS; one instance cannot combine object storage from different cloud providers. MongoDB recommends placing the instance in the same provider region as its data source to avoid excessive charges. (high confidence) (verified as of 2026-10-24) [source]
- Federated sources do not support user-created indexes, change streams or multi-document ACID transactions. MongoDB also documents a 16 MB document limit and says result order requires an explicit sort when ordering matters. (low confidence) (verified as of 2026-10-24) [source]
Facts and statements
- Atlas Data Federation supports Google Cloud Storage buckets as stores and maps their data into virtual collections queried using MongoDB Query Language and supported aggregations. (high confidence) (verified as of 2026-10-14) [source]
- The GCS storage configuration uses stores entries with provider set to gcs, a bucket name, and its region. Virtual collection dataSources refer to a store by storeName and select objects with path; optional prefix and delimiter settings control object traversal. (medium confidence) [source]
- The optional GCS prefix is prepended to a collection data-source path. Path syntax can derive fields from object names; create examples use /json/{collectionName()} with a wildcard collection name to create virtual collections dynamically. (medium confidence) [source]
- Atlas Data Federation supports writing aggregation output to GCS with $out. The GCS query documentation allows buckets with read and write permissions; $out must be the final aggregation stage. This is an export capability, not evidence of general transactional updates to bucket-backed collections. (high confidence) (verified as of 2026-10-14) [source]
- MongoDB limits a federated database instance to object storage from one cloud provider and one account in that provider. Its limitations page also excludes indexes, change streams and multi-document ACID transactions on federated sources. (low confidence) (verified as of 2026-10-18) [source]
- For the GCS permission model, Google distinguishes Storage Object Viewer, which reads and lists objects, from Storage Object Creator, which creates objects but does not permit reading, deleting or overwriting them. These role definitions alone do not establish the complete permissions Atlas requires. (low confidence) (verified as of 2026-10-24) [source]
- As one documented federation client integration, Veezoo connects through the Atlas SQL Interface and configures the federated virtual database and JDBC hostname; this guide does not certify GCS-specific compatibility. (low confidence) (verified as of 2026-10-24) [source]
- The GCS defaultFormat setting supports JSON, BSON, Avro, ORC, TSV, CSV and Parquet, with listed gzip variants for several formats. CSV and TSV require a header row. If defaultFormat is omitted, Data Federation attempts file-type detection from initial bytes. (low confidence) [source]
- Atlas documents Google Cloud service account access for Data Federation resources and requires Project Owner access to configure it. The documented setup prerequisites include an M10+ Atlas cluster hosted on Google Cloud; this is the prerequisite of that setup workflow. (low confidence) (verified as of 2026-10-14) [source]
- Google Cloud defines roles/storage.objectViewer as allowing object reads and listing, while roles/storage.objectAdmin includes object creation and deletion as well. These are different permission scopes; this role reference alone does not establish which exact role Atlas requires for a GCS integration. (low confidence) (verified as of 2026-10-23) [source]
- OneUptime independently publishes a federation walkthrough that names GCS among supported sources, but its configuration, partitioning and output examples use S3. It supplies no GCS-specific implementation test in that walkthrough. (low confidence) [source]
Children
- No children recorded.