Understanding Batch Number Behaviour When Importing Timesheets Using Concurrent API Calls

Summary

When integrating with the Import Timesheets API (/entity/ADVPAU/20.200.001/Timesheet), some integrations use concurrent (multi-threaded) API calls to improve performance. Although the integration correctly uses multiple sessions or tokens for each thread, the system produces duplicate Batch Numbers and inconsistent data during concurrent processing.

Why This Occurs

• The Timesheet Import process is not designed to handle simultaneous inserts or updates efficiently when multiple requests are submitted concurrently

• When multiple API calls are processed concurrently, MYOB Acumatica may allocate the same Batch Number to more than one timesheet.

• This is not related to authentication or session reuse

Recommended Integration Approach

1. Use Sequential API Calls for Timesheet Import

Timesheet imports should be processed sequentially instead of concurrently.
Sequential processing ensures that batch creation occurs safely and prevents Batch Number duplication.

2. Implement Retry Logic

Include retry logic with exponential backoff in your integrations to manage:

• Timeouts

• Throttling

• Transient network conditions

This ensures reliability during sequential execution without causing data duplication

3. Avoid Parallel Timesheet Imports

Concurrent requests, regardless of separate sessions or tokens, can cause batch number collisions.
Avoid multi-threaded or parallel calls for this endpoint.

4. Multiple Sessions or Tokens Are Still Valid

The use of multiple sessions or OAuth tokens is correct overall. However, this approach should not be applied to Timesheet Imports to align with the system's batch handling design