Detecting interruption events within an application workflow

A non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform operations comprising:

• maintaining application state data for a plurality of client devices, wherein the application state data describes current states within an application workflow of a client application operating on the plurality of client devices;
• identifying one or more interruption events at the client application on one or more of the plurality of client devices, wherein each interruption event represents an interruption to the application workflow of the corresponding user;
• identifying states within the application workflow of the client application at times of the interruption events based on the maintained application state data for the client devices;
• logging entries in an interruption log for the interruption events and the identified states, wherein the interruption log comprises a plurality of entries of interruption events and states of application workflows corresponding to those interruption events;
• predicting a number of interruption events for a target state of the application workflow within a time period by applying an interruption prediction model to context data describing the time period and order data describing a plurality of orders placed by a plurality of users of an online system, wherein the interruption prediction model is a machine-learning model that is trained to predict the number of interruption events within time periods based on context data and order data;
• computing an actual number of interruption events for the target state during the time period based on the interruption log;
• comparing the predicted number of interruption events to the actual number of interruption events;
• and
• performing a remedial action based on the comparing of the predicted number of interruption events to the actual number of interruption events.

A system comprising a processor and a non-transitory computer-readable medium, wherein the non-transitory computer-readable medium stores instructions that, when executed by the processor, cause the processor to perform operations comprising:

• maintaining application state data for a plurality of client devices, wherein the application state data describes current states within an application workflow of a client application operating on the plurality of client devices;
• identifying one or more interruption events at the client application on one or more of the plurality of client devices, wherein each interruption event represents an interruption to the application workflow of the corresponding user;
• identifying states within the application workflow of the client application at times of the interruption events based on the maintained application state data for the client devices;
• logging entries in an interruption log for the interruption events and the identified states, wherein the interruption log comprises a plurality of entries of interruption events and states of application workflows corresponding to those interruption events;
• predicting a number of interruption events for a target state of the application workflow within a time period by applying an interruption prediction model to context data describing the time period and order data describing a plurality of orders placed by a plurality of users of an online system, wherein the interruption prediction model is a machine-learning model that is trained to predict the number of interruption events within time periods based on context data and order data;
• computing an actual number of interruption events for the target state during the time period based on the interruption log;
• comparing the predicted number of interruption events to the actual number of interruption events;
• and
• performing a remedial action based on the comparing of the predicted number of interruption events to the actual number of interruption events.