International Business Machines Corporation

Kyndryl, Inc.

Mark David Batinich

Luigi Pichetti

Linwood Earl Loving

Pietro Iannucci

Allison V Pelt

Monitoring and automatically remediating issues that arise at run-time during integrations between Application Program Interfaces (APIs) of two or more endpoint products over an integration framework. The monitoring is facilitated by inserting specialized modules into the integration framework that detects changes in the outputs of the integrated endpoint products and attempts to remedy them by automatically adjusting the output in-transit towards a destination endpoint. The specialized modules can be enhanced by machine learning algorithms trained on previously successful remedies. Remedies may be directed towards schema variations and performance drifts, among others.

Automatic and continuous monitoring and remediation of API integrations

A computer-implemented method includes: receiving, by a computing device, information for composing a first enhanced plug-in, wherein the first enhanced plug-in includes one or more logical components for invoking the first enhanced plug-in by an application running in a different framework as the first enhanced plug-in; receiving, by the computing device, information for creating a workflow that includes a series of a plurality of enhanced plug-ins including the first enhanced plug-in and one or more subsequent second enhanced plug-ins; and invoking, by the computing device, the workflow, wherein the invoking the workflow causes the first enhanced plug-in to receive data from an application running on the same or different framework as the first enhanced plug-in, process the data, and forward the processed data as input data to a particular one of the one or more subsequent second enhanced plug-ins or an endpoint application.

Invoking enhanced plug-ins and creating workflows having a series of enhanced plug-ins

For systems integration, an information published about an application programming interface (API) of a service is parsed to extract a pattern from the information. the pattern is compared with a stored pattern in a repository, wherein the stored pattern corresponds to a known API. When the pattern matches the stored pattern within a threshold degree of match, a conclusion is made that the API of the service is the known API. A collector code module is selected where the collector code module is configured to call the known API. The collector code module is sequenced in an integration sequence, to call the known API. A forwarder code module is also sequence din the integration sequence to forward a data output of the service to a consumer application.

Growing community of inventors

North Aurora, IL, United States of America

Mark David Batinich