<!-- llms-explorer concept facts · https://llms-explorer.com/tree/atlas-live-migration/ · pack 2026-10-23 · ~935 tokens -->

# Atlas Live Migration

> Source-anchored research for Atlas Live Migration. Run status: COMPLETED.

Parent: [MongoDB Migration Patterns](https://llms-explorer.com/tree/mongodb-migration-patterns/) · 4 facets · 8 facts · page: https://llms-explorer.com/tree/atlas-live-migration/

## Definitions

- Atlas Live Migration is a managed Mongosync workflow: it copies source data, continuously applies later source changes, and leaves applications using the source until cutover. AWS Prescriptive Guidance independently describes the same initial-copy, change-monitoring, and cutover sequence. (high confidence) — [source](https://www.mongodb.com/docs/atlas/live-migration.md)

## How it works

- MongoDB's comparison page says to use Atlas Live Migration for source clusters ≤5 TB or ≤3 shards; it calls standalone Mongosync the better option when the source exceeds those stated thresholds and lists tunable parameters, pause and resume, detailed progress tracking, and direct log access for debugging transient errors. The two thresholds are joined by “or,” so the page does not state a combined 5 TB and 3-shard eligibility cap. (low confidence) (verified as of 2026-10-20) — [source](https://www.mongodb.com/docs/atlas/import/live-migration-comparison-modes/)

## Problems, failure modes and limitations

- A January 2025 MongoDB Community Hub thread reports the error `We are unable to track migration progress. Please try again later.` after validation; a respondent associated that symptom with destination oplog exhaustion and suggested increasing destination oplog capacity or disabling embedded verification and verifying manually. (low confidence) — [source](https://www.mongodb.com/community/forums/t/issue-with-live-migration-progress-gcp/309810)

## Facts and statements

- The pull workflow reduces downtime but does not promise an interruption-free switch: its documented cutover stops source writes and application instances, redirects the application to Atlas, and restarts it. AWS guidance also describes changing the application connection string to complete cutover. (high confidence) — [source](https://www.mongodb.com/docs/atlas/live-migration.md)
- Atlas Live Migration does not provide collection or database filtering; standalone Mongosync supports filtered sync. A 2026 RightTouch production write-up independently identifies this gap as the reason to use standalone Mongosync for a collection-only move between Atlas clusters. (high confidence) (verified as of 2026-10-20) — [source](https://www.mongodb.com/docs/atlas/import/live-migration-comparison-modes/)
- For pull live migrations, MongoDB requires the migration lag window to remain within the source cluster's oplog replication window; plan enough source oplog retention for the migration to stay within that bound. (high confidence) (verified as of 2026-10-20) — [source](https://www.mongodb.com/docs/atlas/live-migration/)
- Migration eligibility depends on supported source and destination versions and topologies. MongoDB's current pull-migration matrix lists supported version paths, while the AWS guide scopes its guided pull or push instructions to clusters running MongoDB 6.0.8 or later; verify the exact path against the current Atlas documentation. (high confidence) (verified as of 2026-10-20) — [source](https://www.mongodb.com/docs/atlas/live-migration/)
- Atlas pull live migration requires the source firewall to allow connections from MongoDB's live-migration servers; MongoDB's current instructions also say the pull data path does not use VPC peering or private endpoints, so confirm the network path before choosing this method. (high confidence) (verified as of 2026-10-20) — [source](https://www.mongodb.com/docs/atlas/live-migration/)