Skippr Docs

Appearance

When to Use MongoDB as a Source Connector ​

May 2026

MongoDB is the right source connector when collection documents are the starting point and the team wants a direct path from BSON documents into downstream analytical tables.

Short Answer ​

Use MongoDB as a source connector when the application truth already lives in one collection and the downstream team wants those documents as data, not as a partial export from some intermediate system. It is a good fit when BSON-to-JSON conversion is acceptable, collection access is straightforward, and document-level change capture may matter later.

A common pattern is a product database with users, orders, or events stored as documents. Skippr can read the selected database and collection directly, optionally filter to the subset that matters, and later use MongoDB change streams to keep a warehouse destination aligned to the current document state.

Why Teams Struggle with This ​

MongoDB is often the right source semantically, but teams still need to be deliberate about document identity and downstream shape. A collection can be easy to read and still produce awkward analytical tables if no one decides how the document key should map into rows.

  • The connector depends on a valid MongoDB connection string and collection-level read access.
  • An optional filter is useful for narrowing scope, but it should stay simple enough to remain an understandable integration contract.
  • CDC requires change-stream prerequisites, so batch extraction and real-time change capture are not the same operational commitment.
  • MongoDB is a weak fit when the real need is relational joins at the source instead of document extraction.

How Skippr Handles It ​

Skippr makes MongoDB practical by keeping the source surface small: connection string, database, collection, and optional filter. That is enough to make one collection a stable starting point instead of pushing teams into a bigger migration decision too early.

When the time comes for change streams, the same connector can participate in a final-state CDC path into a supported warehouse destination. That keeps the first useful batch setup and the later CDC setup conceptually aligned.

  • Reads directly from a named MongoDB database and collection.
  • Converts BSON documents to JSON for downstream use.
  • Supports an optional filter document to narrow the extract.
  • Can extend into change-stream CDC for exactly-once final-state destinations.

What the First Useful Version Looks Like ​

The first useful version is one collection, one explicit document key strategy, and one downstream table that proves the document-to-row mapping is understandable.

MongoDB is a bad fit when the team cannot support the connection-string and network requirements, or when the downstream consumers need a relational source contract more than they need the original document model.