The present invention discloses a computer implemented method for developing an anomaly detector which is adapted to detect/predict anomaly in one or more network terminals and optimize the behavior of the network terminals. The said method is adapted to collect and monitor the behavior of the network terminals and compare it with the behavior profile of the network terminals in order to detect the anomaly parameter. The behavior profile is the normal interaction of the software and hardware components of the network terminals. A system for implementation and execution of such anomaly detector is also disclosed.

A method, system and computer program product for detecting potential failures in completing a continuous delivery (CD) pipeline using machine learning. A CD pipeline is defined to include stages, where each stage includes a binary event(s). A model is created by applying an Apriori algorithm and a sequential pattern mining algorithm to a set of previous patterns of sequences of binary events to calculate confidence scores for completing a set of binary events in a particular order. After identifying an ongoing CD sequence (ordered set of binary events) for a software application, the model is used to predict a likelihood of the ongoing CD sequence for the software application completing the CD pipeline by generating confidence score(s) for the ongoing CD sequence. A notification is issued regarding a potential failure in completing the CD pipeline for the software application if a confidence score is below a threshold value.

A method, system and computer program product for detecting potential failures in completing a continuous delivery (CD) pipeline using machine learning. A CD pipeline is defined to include stages, where each stage includes a binary event(s). A model is created by applying an Apriori algorithm and a sequential pattern mining algorithm to a set of previous patterns of sequences of binary events to calculate confidence scores for completing a set of binary events in a particular order. After identifying an ongoing CD sequence (ordered set of binary events) for a software application, the model is used to predict a likelihood of the ongoing CD sequence for the software application completing the CD pipeline by generating confidence score(s) for the ongoing CD sequence. A notification is issued regarding a potential failure in completing the CD pipeline for the software application if a confidence score is below a threshold value.

Technologies for monitoring performance of a machine learning model include receiving, by an unsupervised anomaly detection function, digital time series data for a feature metric; where the feature metric is computed for a feature that is extracted from an online system over a time interval; where the machine learning model is to produce model output that relates to one or more users' use of the online system; using the unsupervised anomaly detection function, detecting anomalies in the digital time series data; labeling a subset of the detected anomalies in response to a deviation of a time-series prediction model from a predicted baseline model exceeding a predicted deviation criterion; creating digital output that identifies the feature as associated with the labeled subset of the detected anomalies; causing, in response to the digital output, a modification of the machine learning model.

Methods, systems and computer program products for automated resolution of computer system incidents are provided.

Detecting data anomalies by receiving a first data set related to a first variable metric, determining data anomaly detection scores for data points of the first data set according to a plurality of data anomaly detection techniques, generating an adaptive ground-truth window according to the data anomaly detection scores, assigning a weighting value to each data point within the adaptive ground-truth window, training a machine learning system using the set of data anomaly detection scores and weighting values, and providing a trained machine learning system for evaluating a second data set.

Various technologies described herein pertain to controlling reconfiguration of a dependency graph for coordinating reconfiguration of a computing device. An operation can be performed at the computing device to detect whether an error exists in the dependency graph for a desired configuration state. The dependency graph for the desired configuration state specifies interdependencies between configurations of a set of features. An error can be detected to exist in the dependency graph when the desired configuration state differs from an actual configuration state of the computing device that results from use of the dependency graph to coordinate configuring the set of features. Feedback concerning success or failure of the dependency graph on the computing device can be sent from the computing device to a configuration source. The dependency graph can be modified (by the computing device and/or the configuration source) based on whether an error is detected in the dependency graph.

A diagnostics tool aids in the efficient collection of relevant error data for addressing faults in connected software systems. Context information is collected from a software system that is being displayed. Configuration of the software system is collected, and used to identify relevant connected software systems. Error data is collected via respective log interfaces from the error logs of the software system being displayed, and relevant connected systems. The context, configuration, and error data is stored in a database. Based at least upon the configuration information, a query is formulated and posed to the database. A corresponding query result is received and processed to return an error report to a user interface, for inspection (e.g., by a user or a support staff member). Certain embodiments may further generate an appropriate recommendation based upon the query result. The recommendation may be generated with reference to a stored ruleset and/or artificial intelligence.

A method for diagnosing and troubleshooting failures of components of a physical system with low troubleshooting cost, according to which for each component in the system, a Model-Based Diagnosis (MBD) is used for computing the probability of causing a system failure, based on currently observed system behavior or on knowledge about the system's structure. Then the probability of causing a system failure is computed, based on its age and its survival curves. Then, it is determined whether a faulty component C should be fixed or replaced by minimizing future troubleshooting costs, being the costs of the process of diagnosing and repairing an observed failure.

A method and a system for diagnosing remaining lifetime of storages in a data center are disclosed. The method includes the steps of: a) sequentially and periodically collecting operating attributes of failed storages along with time-to-fail records of the failed storages in a data center; b) grouping the operating attributes collected at the same time or fallen in a continuous period of time so that each group has the same number of operating attributes; c) sequentially marking a time tag for the groups of operating attributes; d) generating a trend model of remaining lifetime of the storages from the operating attributes and time-to-fail records by ML and/or DL algorithm(s) with the groups of operating attributes and time-to-fail records fed according to the order of the time tags; and e) inputting a set of operating attributes of a currently operating storage into the trend model to calculate a remaining lifetime therefor.