Search code examples
unit-testingautocompleteazure-cognitive-searchautosuggestsearch-suggestion

How can we measure the performance of azure search suggest API for 100 concurrent users?


I have implemented autosuggestion feature in my application and I am using azure search suggest service for providing suggestion. It's working fine for me but I am not sure what will be the impact after making it live when around 100 users will use my application concurrently. Currently I am thinking to use standard(S1) pricing tier.

Is there anyway so that I can test load for 100 users for newly added feature i.e. autocomplete?


Solution

  • Checkout https://learn.microsoft.com/en-us/azure/search/search-traffic-analytics to see how you can add insturmentation for the Search traffic analytics blade to consume. What I would advise is starting at Basic and build a test harness like a console app, that could simulate your 100 concurrent users. There are load test features built into Azure Dev Ops also, you can check out those options at https://learn.microsoft.com/en-us/azure/devops/test/load-test/overview?view=azure-devops. You can then review the data captured in the Search traffic analytics blade to guestimate if Basic will be enough for 100 concurrent users or whether not you need to move up to a Standard teir.