Tools for performance issues
Because performance issues can arise with live sync and the initial sync, a few tools are available to help you diagnose the issues and troubleshoot the problems. The most useful tool is called Trace parser, which you can use for general system performance, including dual-write specific performance issues. From a high-level analysis, as a first step, you can observe the system’s performance or the integration from the Performance timer tool. For more information, see Performance timer.
When you’re finished diagnosing the issues, you can get detailed traces and log analysis by following these links:
- To take a trace in the system, go to Take traces by using Trace parser.
- To diagnose the trace, go to Diagnose issues and analyze performance by using Trace parser.
The trace will provide the SQL call stack to help you determine where the issues are occurring from a data perspective. This tool can help you diagnose if slow queries, business logic implementation, or locks and blocks are present in the data as it’s called from the database.
Trace parser will also provide you with the activity ID of the process, which you can use in the Lifecycle Services monitoring tools to check the logs and filter them with the ProviderName = Microsoft-Dynamics-Ax-DualWriteSync command.
Along with diagnosing with tools like Trace parser, key bottlenecks for performance also exist. If too many business events are raised, you can check the events from Lifecycle Services by using the Activity ID, the TaskName = DualWriteRecordsCached command, and the ProviderName = Microsoft-Dynamics-AX-BusinessEventsRuntime command. Multiple maps that share the same data sources can trigger updates that might cause too many business events and updates.
Slow database queries can also contribute to the overall system processes and performance. You can identify these queries from the Lifecycle Services monitoring and activity log by filtering on the TaskName = AosDatabaseSlowQuery command. This command will show each query from the system and will allow you to view the milliseconds, and it will determine which queries are the slowest.
Testing the entities for performance can be helpful. You can complete this task by inserting and updating the records by using OData calls from any external tools, such as the Excel Add-in. Testing the performance in this way can determine the cause of the performance issues. Determining whether the performance is related to dual-write or if it’s related to the entity is an important distinction.