This document summarizes the configuration we used to perform load-testing for the ChattyIO anti-pattern. You should also read about our general approach to deployment and load testing.
| Option | Value |
|---|---|
| Compute | Web App |
| Tier | P1 |
| Instance Count | 1 |
| SQL Tier | P3 |
| Max Pool Size | 1000 |
The load test project included two webtests, each invoking an HTTP GET operation.
The URLs used were:
- http://yourwebapp.azurewebsites.net/chattyproduct/products/{SubcategoryId}
- http://yourwebapp.azurewebsites.net/chunkyproduct/products/{SubcategoryId}
Replace yourwebapp with the name of your web application, and replace {SubcategoryId} with subcategory number in the range 1 to 5 generated by using the Generate Random Integer plugin.
The project also included two load tests, one for each web test. Both load tests were run against a single deployment but at different times, using the following parameters:
| Parameter | Value |
|---|---|
| Initial User Count | 1 |
| Maximum User Count | 1000 |
| Step Duration | 60s |
| Step Ramp Time | 0s |
| Step User Count | 100 |
| Test Duration | 15 minutes |
| Test Warm Up | 30 seconds |
The load test for the http://yourwebapp.azurewebsites.net/chattyproduct/products/{SubcategoryId} web test generated the following results:
The load test for the http://yourwebapp.azurewebsites.net/chunkyproduct/products/{SubcategoryId} web test generated the following results:
