公开(公告)号：US20230093659A1

公开(公告)日：2023-03-23

申请号：US17479518

申请日：2021-09-20

Applicant: salesforce.com, inc.

Inventor： Bobert M. O'DELL ,  Nicolas Hernan BATTIATO ,  Diego Gabriel LARRALDE ,  Guido Agustin MARTINEZ ,  Christian VALLEJOS ,  Maria Florencia VIMBERG ,  Eduardo Cominguez ,  Ignacia Agustin Manzano ,  Peter Gorski

IPC: G06F11/34 ,  G06F9/54

Abstract: Disclosed herein are system, method, and computer-readable medium embodiments for providing the ability to automate identification of endpoints of an API for potential health checks based on machine learning and/or similarity search algorithms. The algorithms analyze an API specification according to the algorithm's training and/or search among reference APIs. Rather than having to manually generate health check tests for the web service, a test developer can interact with a test service through a web browser and provide the service an API specification. The test service then can automatically rank identified endpoints according to a ranking system as well as automate health checks by automatically preparing the health check code according to each identified endpoint.

公开(公告)号：US20230089336A1

公开(公告)日：2023-03-23

申请号：US17479493

申请日：2021-09-20

Applicant: salesforce.com, inc.

Inventor： Nicolas BATTIATO ,  Diego Gabriel LARRALDE ,  Guido MARTINEZ ,  Christian VALLEJOS

IPC: G06F11/36 ,  G06F11/34 ,  G06F11/30 ,  G06F11/32

Abstract: Disclosed herein are system, method, and computer-readable medium embodiments for providing the ability to automate the process of generating load tests used for benchmarking APIs. Rather than having to manually generate load tests for a web service API, a test developer can interact with a test service through a web browser and provide the service an API specification and testing parameters. The test service can analyze the API specification, automatically identified endpoints of the API, and generate load tests according to the expected input/output structures of the endpoints. The automatic load test generation can proceed by referring to a library of test instructions and extracting the portions of the test instructions that work for the identified endpoints.