Abstract: In a “buy online, pick up in-store” service, customers place orders for items that are retrieved from store inventory and packaged for easy pick-up by the customer. Since these services typically fulfill orders from current store inventory, some items purchased by users are unavailable at the time of order fulfillment. In these circumstances, a recommendation system identifies a recommended replacement item using a trained model. The trained model includes a hierarchy of multiple sub-models, where each sub-model is configured to receive a different set of features of items as input and to generate, as output, a candidate recommended replacement item. A recommended replacement item is selected from the candidate recommendations generated by the multiple sub-models and sent for display to a user. The recommendation system receives user feedback regarding the recommended replacement item and selectively retrains one or more sub-models of the multiple sub-models based on the user feedback.

Abstract: A method for maintaining applications may include: (1) receiving a request to recover a first application, (2) identifying a first production topology of the first application that identifies a set of resources upon which the application depends, (3) maintaining a template for transforming the first production topology of the first application into a first recovery topology for the first application, the template comprising information for mapping the first production topology to the first recovery topology, (4) applying the template to the first production topology at a first point in time to create the first recovery topology, and (5) recovering the first application to a first computing system using the first recovery topology. Various other methods, systems, and computer-readable media are also disclosed herein.

Abstract: A method for maintaining applications may include: (1) receiving a request to recover a first application, (2) identifying a first production topology of the first application that identifies a set of resources upon which the application depends, (3) maintaining a template for transforming the first production topology of the first application into a first recovery topology for the first application, the template comprising information for mapping the first production topology to the first recovery topology, (4) applying the template to the first production topology at a first point in time to create the first recovery topology, and (5) recovering the first application to a first computing system using the first recovery topology. Various other methods, systems, and computer-readable media are also disclosed herein.

Abstract: Techniques are disclosed relating to storing a log of write operations made to a first storage device by one of a plurality of host computers running an instance of a distributed application. The log of write operations is stored at a second storage device. The plurality of host computers communicate status information to the second storage device over respective communication paths. Upon a failure to communicate status information between one of the host computers and the second storage device, the second storage device reads from a predetermined location in the first storage device to determine whether the host computer is still performing write operations. If the second storage device reads an expected signature value written by the host computer, the host computer is deemed to have written data, which indicates that the host computer is operational but that the write operations have not been recorded by the second storage device.