Automating the configuration and tuning of a disaster recovery preparation system

A computer-implemented method for tuning a disaster recovery preparation system for a storage system comprising:

• receiving a trigger associated with a recovery point objective that is not met and a state of the storage system;
• inputting one or more of the trigger or the state into a machine learning agent comprising a machine learning model, wherein the machine learning agent outputs an action based on the state of the storage system and the recovery point objective that is not met, wherein the machine learning agent comprises an action space that includes at least two actions of:
• adjusting a time window when disaster recovery preparation operations are performed;
• changing a bandwidth allocated to the disaster recovery preparation system;
• updating a priority assigned to the disaster recovery preparation system;
• or
• changing a number of snapshots that can be concurrently performed and replicated by the disaster recovery preparation system, and the action includes at least one of the at least two actions;
• performing, in the disaster recovery preparation system, the action to change the state of the storage system;
• after performing the action, monitoring one or more metrics of the disaster recovery preparation system or the storage system;
• and
• training the machine learning model based on the monitoring by tuning one or more parameters of the machine learning model based on the one or more metrics.

A system comprising:

• a memory that stores instructions, and one or more processors that are coupled to the memory and, when executing the instructions, is configured to:
• receive a trigger associated with a recovery point objective of a disaster recovery preparation system that is not met:
• determine a state of a storage system;
• input one or more of the trigger and the state into a machine learning agent comprising a machine learning model, wherein the machine learning agent outputs an action based on the state of the storage system and the recovery point objective that is not met, wherein the action is selected from machine learning agent comprises an action space of the machine learning agent that includes at least two actions of:
• adjusting a time window when disaster recovery preparation operations are performed;
• changing a bandwidth allocated to the disaster recovery preparation system;
• updating a priority assigned to the disaster recovery preparation system;
• or
• changing a number of snapshots that can be concurrently performed and replicated by the disaster recovery preparation system, and the action includes at least one of the at least two actions;
• perform, in the disaster recovery preparation system, the action to change the state of the storage system;
• after performing the action, monitor one or more metrics of the disaster recovery preparation system or the storage system;
• and
• train the machine learning model based on the monitoring by tuning one or more parameters of the machine learning model based on the one or more metrics.