Abstract: A computer-implemented method is presented for detecting anomalies in dynamic datasets generated in a cloud computing environment. The method includes monitoring a plurality of cloud servers receiving a plurality of data points, employing a two-level clustering training module to generate micro-clusters from the plurality of data points, each of the micro-clusters representing a set of original data from the plurality of data points, employing a detecting module to detect normal data points, abnormal data points, and unknown data points from the plurality of data points via a detection model, employing an evolving module using a different evolving mechanism for each of the normal, abnormal, and unknown data points to evolve the detection model, and generating a system report displayed on a user interface, the system report summarizing the micro-cluster information.

Abstract: Embodiments facilitating performance anomaly detection are described. A computer-implemented method comprises: detecting, by a device operatively coupled to one or more processing units, based on monitoring data of a plurality of performance metrics of a monitored device, at least one trend within the monitoring data of the respective performance metrics; removing, by the device, the at least one trend from the monitoring data of the respective performance metrics to generate modified data of the respective performance metrics; and detecting, by the device, a performance anomaly based on the modified data of the respective performance metrics and a behavior clustering model comprising at least one steady state.

Abstract: Embodiments are directed to a computer implemented method for generating a drift detector. The method includes generating, using a processor system, drift cases based at least in part on known drift set data of a computer system. The method further includes injecting, using the processor system, the drift cases into the computer system to generate a first data set. The method further includes applying, using the processor system, cleaning rules to the first data set to reduce a size of the first data set and generate a cleaned data set. The method further includes extracting one or more features of the cleaned data set. The method further includes normalizing the extracted one or more features of the cleaned data set. The method further includes training a machine learning system using the extracted and normalized one or more features of the cleaned data, wherein an output of the machine learning system comprises the drift detector.

Abstract: A computer-implemented method is presented for detecting anomalies in dynamic datasets generated in a cloud computing environment. The method includes monitoring a plurality of cloud servers receiving a plurality of data points, employing a two-level clustering training module to generate micro-clusters from the plurality of data points, each of the micro-clusters representing a set of original data from the plurality of data points, employing a detecting module to detect normal data points, abnormal data points, and unknown data points from the plurality of data points via a detection model, employing an evolving module using a different evolving mechanism for each of the normal, abnormal, and unknown data points to evolve the detection model, and generating a system report displayed on a user interface, the system report summarizing the micro-cluster information.

Abstract: Techniques for Bayesian-based event grouping are provided. One technique includes determining a group of alarm events from received alarm events; in response to the group of alarm events matching a group of historical alarm events, determining a first correlation, wherein the group of historical alarm events comprises correlated events associated with a same entity; and determining a root cause of the group of alarm events based on the first correlation.

Abstract: Techniques that facilitate identification and/or analysis of sequences associated with computing devices are provided. In one example, a system includes a transaction component, a clustering component and a model component. The transaction component identifies at least one sequence in a stream of sequences generated by a computing device in communication with the system. The clustering component assigns the at least one sequence to a transaction sequence group. The model component generates a transaction model based on the transaction sequence group.

Abstract: Embodiments facilitating performance anomaly detection are described. A computer-implemented method comprises: detecting, by a device operatively coupled to one or more processing units, based on monitoring data of a plurality of performance metrics of a monitored device, at least one trend within the monitoring data of the respective performance metrics; removing, by the device, the at least one trend from the monitoring data of the respective performance metrics to generate modified data of the respective performance metrics; and detecting, by the device, a performance anomaly based on the modified data of the respective performance metrics and a behavior clustering model comprising at least one steady state.

Abstract: A computer-implemented method is presented for automatically generating alerting rules. The method includes identifying, via offline analytics, abnormal patterns and normal patterns from history logs based on machine learning, statistical analysis and deep learning, the history logs stored in a history log database, automatically generating the alerting rules based on the identified abnormal and normal patterns, and transmitting the alerting rules to an alerting engine for evaluation. The method further includes receiving a plurality of online log messages from a plurality of computing devices connected to a network, augmenting the plurality of online log messages, and extracting information from the plurality of augmented online log messages to be provided to the alerting engine, the alerting engine configured to approve and enforce the alerting rules automatically generated by the offline analytics processing.

Abstract: A computer-implemented method is presented for detecting anomalies in dynamic datasets generated in a cloud computing environment. The method includes monitoring a plurality of cloud servers receiving a plurality of data points, employing a two-level clustering training module to generate micro-clusters from the plurality of data points, each of the micro-clusters representing a set of original data from the plurality of data points, employing a detecting module to detect normal data points, abnormal data points, and unknown data points from the plurality of data points via a detection model, employing an evolving module using a different evolving mechanism for each of the normal, abnormal, and unknown data points to evolve the detection model, and generating a system report displayed on a user interface, the system report summarizing the micro-cluster information.

Abstract: Techniques for log-based diagnosis for declarative-deployed applications can comprise: based on a labeled deployment declaration, classifying, by a device operatively coupled to a processor, a runtime log associated with an application, resulting in a classified runtime log. Techniques can also comprise: based on the classified runtime log, training, by the device, an aggregation model to represent a defined state of the application.

Abstract: Systems and methods for ticket classification and response include labeling tickets with a ticket classifier that assigns a ticket label and an associated confidence score to each ticket. Tickets are clustered according to semantic similarity to form ticket clusters. A template associated with each ticket cluster is determined. Templates and the respective ticket clusters are clustered according to semantic similarity to form one or more ticket super-clusters. Tickets that have below-threshold confidence scores are labeled according to the one or more ticket super-clusters. The tickets are automatically responded to.

Abstract: Methods and systems for ticket classification and response include clustering tickets according to semantic similarity to form ticket clusters. A template associated with each ticket cluster is determined that includes an invariant portion and a variable portion. A new ticket sub-class, based on the variable portion of the template, is determined that represents a specific sub-type of an existing class. A ticket taxonomy is updated to include the new ticket sub-class. The tickets are labeled according to the updated ticket taxonomy. The tickets are automatically responded to.

Abstract: A method or apparatus for monitoring a system by detecting misbehaving components in the system is presented. A computing device receives historical data points based on a set of monitored signals of a system. The system has components that are monitored through the set of monitored signals. For each monitored component, the computing device performs unsupervised machine learning based on the historical data points to identify expected states and state transitions for the component. The computing device identifies one or more steady components based on the identified states of the monitored components. The computing device also receives real-time data points based on monitoring the set of signals from the system. For each identified steady component, the computing device examines the received real-time data points for deviation from the expected state and state transitions of the steady component. The computing device reports anomaly in the system based on the detected deviations.

Abstract: A method for automatically detecting and diagnosing problems in computer system functioning includes determining changed objects from computer system monitoring data, calculating temporal correlations from errors and changes sequences for each changed object, identifying and ranking suspicious computer system behavior patterns from the temporal correlations, and outputting said ranked suspicious computer system behavior patterns.

Abstract: Techniques for log-based diagnosis for declarative-deployed applications can comprise: based on a labeled deployment declaration, classifying, by a device operatively coupled to a processor, a runtime log associated with an application, resulting in a classified runtime log. Techniques can also comprise: based on the classified runtime log, training, by the device, an aggregation model to represent a defined state of the application.

Abstract: A method or apparatus for monitoring a system by detecting misbehaving components in the system is presented. A computing device receives historical data points based on a set of monitored signals of a system. The system has components that are monitored through the set of monitored signals. For each monitored component, the computing device performs unsupervised machine learning based on the historical data points to identify expected states and state transitions for the component. The computing device identifies one or more steady components based on the identified states of the monitored components. The computing device also receives real-time data points based on monitoring the set of signals from the system. For each identified steady component, the computing device examines the received real-time data points for deviation from the expected state and state transitions of the steady component. The computing device reports anomaly in the system based on the detected deviations.

Abstract: A computer-implemented method obtains an action sequence that includes a plurality of actions executed on behalf of a plurality of users for achieving at least one goal. An event sequence that includes a plurality of events associated with types of the plurality of actions is generated from the obtained action sequence. An association model based on the generated event sequence is determined. The association model defines a chronological relationship among events associated with the at least one goal.

Abstract: The scope of the system changes to be considered for analysis for finding problematic changes is reduced in order to allow focusing on highly potential suspicious drifts caused by change sequences. The method and system includes a data cleaning module to remove irrelevant changes, a feature extraction and normalization module to extract the features of change objects, data annotation module to remove irrelevant changes based on patterns, and a clustering module to obtain groups for further analysis. Data cleaning is simplified using domain independent rules. Additional sources of change sequences are removed by application of pattern based techniques so as to narrow down problematic system changes to analyze for root cause analysis. Change error sequence and degree of temporal correlation to correlate system changes with errors, as well as change behavior patterns may be used for downsizing the diagnosis scope.

Abstract: Determining a characteristic of a configuration file that is used to discover configuration files in a target machine, a computer identifies, using information associated with a configuration item of a machine, a candidate configuration file related to the configuration item of the machine, from among a plurality of files from the machine. The computer extracts a value of a feature of the candidate configuration file and aggregates the candidate configuration file with a second candidate configuration file related to the same configuration item identified from among a plurality of files from a second machine, based on the extracted value. The computer then determines a configuration file related to the configuration item from among the aggregated candidate configuration files based on a result of the aggregation, and determines a characteristic of the configuration file related to the configuration item.

Abstract: A computer-implemented method obtains an action sequence that includes a plurality of actions executed on behalf of a plurality of users for achieving at least one goal. An event sequence that includes a plurality of events associated with types of the plurality of actions is generated from the obtained action sequence. An association model based on the generated event sequence is determined. The association model defines a chronological relationship among events associated with the at least one goal.