I have a database used by a multi-tenant application. Recently we are having some performance issues with the database. I think adding some proper indexes will fix this issue. I read about Azure SQL Advisor. Microsoft says they are analyzing usage patterns and provide index recommendations using a machine learning algorithm. I feel like I can use the profiler to trace the workload and use the sql tuning advisor to get the index recommendations. What is the advantage of using Azure SQL advisor which uses a machine learning algorithm? What is the big deal about it?
A machine learning algorithm (here the advisor) takes data as an input (here the usage patterns) and build a mathematical model from it. This model a generalization of the data. An application of such a program could be to estimate a future event or optimising a behaviour (by finding the minimum or maximum of a mathematical functional). Here the index recommendation is optimised.
Of course you might be able to fix the issue by yourself. If you are an expert on the issue and know what to do you don't need the machine learning algorithm. The algorithm is an alternative that is doing the job for you. Who has the better results depends on the input data and the algorithm on the one hand and your expertise on the other hand.