mongodb-transactions_skill

Overview

This skill teaches MongoDB ACID transactions for multi-document operations using Python patterns and practical examples. It covers session lifecycle, transaction options (read/write concerns and read preference), error handling and retry strategies, and production considerations to guarantee data consistency. Learn real-world patterns like money transfers, order placement, and account reconciliation with clear constraints and limits.

How this skill works

The skill walks through creating and managing sessions, starting and committing or aborting transactions, and running operations with the session context so multiple collection changes are atomic. It explains transaction options such as snapshot readConcern, majority writeConcern, and readPreference, and shows how to detect transient errors and implement retries. Examples demonstrate common flows, resource cleanup (endSession), and how transactions roll back on failure.

When to use it

• When multiple documents or collections must be updated atomically (e.g., money transfer).
• Order processing that updates orders, inventory, and payment records together.
• Account reconciliation where sums and deletions must remain consistent.
• When business logic requires all-or-nothing guarantees across collections.
• Only on replica sets or supported sharded clusters (not standalone/shared free clusters).

Best practices

• Keep transactions short to reduce lock contention and latency.
• Use appropriate readConcern and writeConcern (snapshot + majority) for correctness.
• Order operations consistently to reduce deadlock risk.
• Implement retry logic for TransientTransactionError and handle UnknownTransactionCommitResult specially.
• Avoid heavy write volumes inside a single transaction (observe 16MB limits).

Example use cases

• Transfer funds between accounts: two updates inside one transaction to prevent partial transfers.
• Place an order: create order, decrement inventory, and charge user within one transaction.
• Reconcile accounts: aggregate balances, update main account, delete secondaries, and log the action atomically.
• Inventory adjustments where multiple product documents must remain consistent after a sale.