Skippr Docs

Appearance

What Network Access Skippr Needs for MongoDB ​

June 2026

MongoDB source access starts with a reachable connection string and a user that can read the target database and collection.

Short Answer ​

Skippr needs a MongoDB connection string that the runner can actually use, plus read access to the selected database and collection. The connector docs define the source around connection_string, database, collection, and optional filter, so the useful boundary is the exact endpoint and auth context embedded in that URI, not just the idea that MongoDB exists somewhere inside the network.

That becomes concrete fast. If the source is mydb.users, the runner needs to resolve the host in the connection string, reach the MongoDB endpoint from its own network, authenticate as a user that can read mydb.users, and return documents from that collection. If CDC is part of the plan, the access story expands again because change streams have their own prerequisites beyond ordinary batch reads.

Why Teams Struggle with This ​

MongoDB connections can look healthy while the actual source boundary is wrong. Teams often validate the URI with one admin shell, then run Skippr from a different network segment or with a user that can connect but cannot read the intended collection.

  • The connection string has to be valid for the runner environment, not just for a developer laptop.
  • The MongoDB user needs read access to the selected database and collection.
  • A wrong database or collection name often looks like an empty extract rather than an obvious connection failure.
  • CDC through change streams adds requirements beyond the basic batch-read path.

How Skippr Handles It ​

Skippr keeps the source definition focused on the real MongoDB boundary: one connection string, one database, one collection, optional filter. That makes it easier to reason about reachability because the connector is attached to one concrete slice of MongoDB instead of a broad server-level promise.

It also helps with troubleshooting language. The docs say the runner must be able to reach the MongoDB endpoint and the user needs read access, so the operational question becomes whether this exact URI can read this exact collection from the machine that is running the sync.

  • Uses a standard MongoDB connection URI as the network and auth boundary.
  • Targets one database and one collection explicitly.
  • Supports optional JSON filtering when the first validation should stay narrow.
  • Leaves CDC-specific change-stream requirements as a deliberate next step instead of assuming them.

What the First Useful Version Looks Like ​

The first useful version is one collection with a known document shape, a runner-hosted environment variable for the connection string, and a small extract that proves the URI works where the job will really run.

If the read path is inconsistent across environments, solve that first. Nothing about downstream modeling helps until the collection is reliably reachable.