Computing technology using artificial intelligence/machine learning methods and systems for a storage system is provided. One method includes measuring by a processor, a first metric indicating health of a first storage system in storing and retrieving data; quantifying by the processor, the first metric based on comparison of the first metric with a same metric for a plurality of storage systems; identifying by the processor, a plurality of features potentially impacting the first metric, based on a predictive, machine-learning algorithm built on performance and configuration data for the plurality of storage systems; selecting, a first feature by the processor, based on impact of the first feature on the first metric; and generating a command by the processor for making a change to the first feature.