Atlas Data API Removal
Parent: MongoDB BI Connector and SQL Access · researched 2026-05-28T16:44:56.847Z· 5 sources · 6 concepts · skill mongodb-bi-connector
Two distinct approaches for SQL-based BI tool integration with MongoDB:
Overview
- Two distinct approaches for SQL-based BI tool integration with MongoDB: [source]
- Legacy BI Connector (mongosqld): EOL September 2026. Self-hosted translation layer. [source]
- Atlas SQL Interface (current): Managed service via Atlas Data Federation. Certified connectors for Tableau and Power BI. [source]
What It Is
- A self-hosted binary (mongosqld) that acts as a MySQL-protocol SQL translation layer in front of MongoDB. BI tools connect via ODBC/JDBC as if connecting to a MySQL database. [source]
DRDL Schema Files
- mongosqld uses DRDL (Document Relational Definition Language) files to define the SQL schema mapping: [source]
EOL Timeline
Architecture
- Atlas SQL Interface uses Atlas Data Federation (FDI) as its query engine. No self-hosted binary needed. [source]
Certified Connectors
- Custom connector in the Tableau Exchange or direct download [source]
- Supports Tableau Desktop, Tableau Server, Tableau Prep [source]
- Tableau Cloud: check current status (in progress as of 2025) [source]
- Driver: JDBC or MongoDB ODBC Driver (Windows/Linux/macOS) [source]
- Certified by Microsoft; available in Microsoft AppSource [source]
- Supports DirectQuery (live Atlas queries) and Import mode [source]
- Driver: MongoDB ODBC Driver [source]
- Via ODBC connection with MongoDB ODBC Driver [source]
- Windows only (ODBC data source configuration) [source]
JDBC Connection
ODBC Connection (Power BI / Excel)
DirectQuery vs Import Mode (Power BI)
- DirectQuery performance depends on Atlas query performance. Index optimization applies here too - slow MQL = slow DirectQuery. [source]
Atlas Cluster Requirements
Cost
- Atlas SQL Interface incurs Data Federation query processing costs: [source]
- $5.00/TB of data processed per query [source]
- Partition strategy and Parquet format reduce costs significantly [source]
- DirectQuery workloads (multiple users, frequent refreshes) accumulate more data processed charges than scheduled Import [source]
Pre-migration checklist
Schema Migration
- DRDL columns → jsonSchema properties: [source]
Anti-Patterns
- Using BI Connector for new deployments (post-2024): EOL is September 2026; invest in Atlas SQL Interface [source]
- DirectQuery without optimizing MongoDB indexes: Every Power BI interaction queries Atlas live; unindexed queries = slow dashboards [source]
- Schema auto-infer on heterogeneous collections: Auto-infer samples 1,000 documents and may miss fields in outliers; validate schema explicitly for production [source]
- Not setting a Data Federation query byte limit: Runaway DirectQuery workloads can drive significant $5/TB costs [source]
References
Children
- Express MongoDB Driver Pattern (frontier)
- FastAPI Motor Pattern (frontier)
- Lambda MongoDB Driver Pattern (frontier)
- Delbridge Data API (frontier)
- RESTHeart (frontier)
- Atlas GraphQL API Removal (frontier)
Frontier under this node: Delbridge Data API, Express MongoDB Driver Pattern, FastAPI Motor Pattern, Lambda MongoDB Driver Pattern, RESTHeart