“Looking inside” RetailServer/CRT/SQL with Commerce Runtime Analyzer

This post originally appeared on https://dynamicsnotes.com/looking-inside-retailserver-crt-sql-with-commerce-runtime-analyzer/.

The Retail Product team built a tool that can be used to trace the RetailServer calls, CRT executes, and SQL calls. It shows call-stack like information based on event traces, similarly to the Dynamics 365 Trace Parser. The tool can be used to find slow handlers, slow SQL or just understand what is happening as part of a call.

As usual, a picture says more than 100 words. See my investigation of slow performance of the UpdateCartLines RetailServer call. From that call stack, I can learn this flow includes a lot of other internal steps: get identity, get cart, get products, get unit of measures, calculate reason codes, calculate prices tax, calculate discounts, calculate sales tax, calculate totals, save versioned cart, final saving of the transaction in SQL and much more. This is VERY valuable information and can be obtained with the Commerce Runtime Analyzer WITHOUT reflecting the source code.


I notice that the CalculateCartRequiredReasonCodesServiceRequest takes 2021 ms. When I add up the included child timings, there is a full 2s missing. This is in my own customization (for this sample, I have added a 2s delay in an extended handler).

Another scenario is to find out what data the CRT logic is using. I ran a product search and the tool lets me know that the data for the 62 result rows comes from [crt].[GetProductsSearchResultsByText_V2].

The tool is not perfect. At this time, it does not seem to show custom handlers properly. If the Retail team gets enough “requests”, I am sure this could be an easy addition made.

In order to use this tool follow the steps:

1. Ask me for a link of to CommercePerfAnalyzer.zip
2. Unzip and execute the CommercePerfAnalyzer.exe
3. Exercise your scenario
4. Wait a few seconds
5. Inspect the output of the tool

One thought on ““Looking inside” RetailServer/CRT/SQL with Commerce Runtime Analyzer”

Leave a Reply

Your email address will not be published. Required fields are marked *