Recommending changes to a storage system

A non-transitory computer readable storage medium storing instructions which, when executed, cause a processing device to:

• generate a plurality of load models that predict corresponding performance loads of a storage system based on characteristics of workloads executing on the storage system, wherein the plurality of load models corresponds to respective configurations of the storage system and wherein the storage system comprises a plurality of storage controllers operatively connected to a plurality of solid state storage devices;
• generate predicted characteristics for one or more workloads executed by the storage system using telemetry data associated with the one or more workloads, wherein generating the predicted characteristics comprises performing a time-series analysis of the one or more workloads;
• predict the corresponding performance loads on the storage system in dependence upon the plurality of load models and the predicted characteristics of the one or more workloads;
• identify one or more of the respective configurations that cause one or more metrics of the storage system to remain within an acceptable range for a determined amount of time based on the predictions of the plurality of load models;
• select, from the one or more of the respective configurations, a particular configuration that improves operation of the storage system;
• present, on a graphical user interface, a recommendation that includes the particular configuration;
• receive information describing one or more modifications to the storage system;
• generate a plurality of subsequent load models of the modified storage system;
• and
• predict an updated performance load using one or more of the plurality of subsequent load models of the modified storage system.

An apparatus comprising:

• a memory;
• and
• a processing device, operatively coupled to the memory, configured to:
• generate a plurality of load models that predict corresponding performance loads of a storage system based on characteristics of workloads executing on the storage system, wherein the plurality of load models corresponds to respective configurations of the storage system and wherein the storage system comprises a plurality of storage controllers operatively connected to a plurality of solid state storage devices;
• generate predicted characteristics for one or more workloads executed by the storage system using telemetry data associated with the one or more workloads, wherein generating the predicted characteristics comprises performing a time-series analysis of the one or more workloads;
• predict the corresponding performance loads on the storage system in dependence upon the plurality of load models and the predicted characteristics of the one or more workloads;
• identify one or more of the respective configurations that cause one or more metrics of the storage system to remain within an acceptable range for a determined amount of time based on the predictions of the plurality of load models;
• select, from the one or more of the respective configurations, a particular configuration that improves operation of the storage system;
• present, on a graphical user interface, a recommendation that includes the particular configuration;
• receive information describing one or more modifications to the storage system;
• generate a plurality of subsequent load models of the modified storage system;
• and
• predict an updated performance load using one or more of the plurality of subsequent load models of the modified storage system.