Harness Inc.

Cisco Technology, Inc.

Appdynamics, Inc.

Raghvendra Singh

Jyoti Bansal

Ido Carmel

Takashi Arai

Andrew Morgan

Sriram Parthasarathy

Parnian Zargham

Rishikesh Singh

Abhijit Sawant

Bhaskar Sunkara

Sastry Konduri

Abhi Sawant

A system can monitor applications and analyze the metrics to determine if one or more of the applications are regressing or performing as expected. The metric analysis includes performing a first short term data analysis and, if data is not as expected, a second short term analysis based on machine learning-based pattern recognition machines. If the short-term analysis finds the metrics aren't as expected, a long-term analysis is performed. The long-term analysis can compare chunks of streaming metric data to cached metric blocks and historical data, and can include a concept drift analysis.

Continuous system service monitoring using real-time short-term and long-term analysis techniques

The present system provides continuous delivery and service regression detection in real time based on log data. The log data is clustered based on textual and contextual similarity and can serve as an indicator for the behavior of a service or application. The clusters can be augmented with the frequency distribution of its occurrences bucketed at a temporal level. Collectively, the textual and contextual similarity clusters serve as a strong signature (e.g., learned representation) of the current service date and a strong indicator for predicting future behavior. Machine learning techniques are used to generate a signature from log data to represent the current state and predict the future behavior of the service at any instant in time.

Service regression detection using real-time anomaly detection of log data

A system monitors applications, analyzes metrics, and provides a dashboard that communicates whether an application is performing as expected. The metric analysis includes performing one or more of a first short term data analysis, a second short term analysis based on machine learning-based pattern recognition machines, and a long-term analysis is performed. Transaction performance metrices are determined based on the monitored of the application. The transaction performance metrices are scored, scaled, and aggregated into a single scaled representation for the application. The scaled application value is then reported to a user through a dynamically updated dashboard. The dashboard displays graphical information representing the health of monitored transactions over time. The reported information can be expanded to additional layers of detail.

Reporting continuous system service status based on using real-time short-term and long-term analysis techniques

Business transactions and the nodes processing the transactions may be monitored. Actions may be applied to one or more nodes when a performance issue is detected. A performance issue may relate to a metric associated with a transaction or node that processes the transaction. If a performance metric does not satisfy a health rule, the policy determines which action should be performed to correct the performance of the node. The corrective action may be applied to a node other than the node at which the performance metric is associated with. For example, if a performance metric for a first node does not satisfy a threshold, the corrective action may be applied to a second node. When a solution applied to a second node is found to correct the problem in first node, the solution may be applied to the other nodes experiencing the same problem.

Server performance correction using remote server actions

The present system uses delegates installed in remote environments to called and transmit, to a remote manager, time series metric data (or data from which metrics can be determined) in real-time. The numerical time series data is persisted, and a learned representation is generated from the data, for example by discretization. The learned representation is then clustered, the clusters are compared to new data, anomalies are determined, and deviation scores are calculated for the anomalies. The derivation scores are compared to thresholds, and results are reported through, for example, a user interface, dashboard, and/or other mechanism.

Service regression detection using real-time anomaly detection of application performance metrics

Business transactions and the nodes processing the transactions are monitored and actions are applied to one or more nodes when a performance issue is detected. A performance issue may relate to a metric associated with a transaction or node that processes the transaction. If a performance metric determined from data captured by monitoring does not satisfy a health rule, the policy determines which action should be performed to correct the performance of the node. When a problem is detected for multiple nodes, the present technology may address a subset of the multiple nodes rather than apply an action to each node experiencing the problem. When a solution is found to correct the problem with the subset of nodes, the solution may be applied to the other nodes experiencing the same problem.

Growing community of inventors

Fremont, CA, United States of America

Raghvendra Singh