Back to results
US 10,884,805 B2Granted
Dynamically configurable operation information collection
AssigneeAmazon Technologies, Inc.
InventorsWei Huang, Matt Nakoneczny
Life of this patent
Estimating the term…
Abstract
Systems and methods are described for the collection and transmission of virtual machine resource operations information. Individual agents on virtual machine resources collect and store operations information in accordance with a current operations information collection configuration. The individual agents will initiate a transmission of the collected operations information. Responsive to the receipt of the transmission of the collected operations information, the monitoring processing service calculates a hierarchy of anomaly scores utilizing machine learning techniques. The monitoring processing service can generate a processing result.
Claims20
A computer-implemented method to manage the processing of monitoring information by virtual machine resources comprising:
- obtaining, at a metric processing component, operations information, the operations information including at least one collected metric attribute by an agent associated with a virtual machine resource;
- generating, by the metric processing component, a hierarchical structure made up of a set of anomaly scores for individual levels of the hierarchical structure, wherein the set of anomaly scores includes one or more group level anomaly scores formed from the obtained operations information, one or more component level anomaly scores formed from the one or more group level anomaly scores, and a system level anomaly score formed from the one or more component level anomaly scores, wherein the set of anomaly scores is determined based on applying a machine learning technique;
- identifying, by the metric processing component, the system level anomaly score indicative of an anomaly;
- identifying, by the metric processing component, a component level anomaly score that is determined to be a source of the identified system level anomaly score wherein the identified system level anomaly score is higher in the hierarchical structure than the identified component level anomaly score;
- identifying, by the metric processing component, a group level anomaly score that is determined to be a source of the identified component level anomaly score, wherein the identified component level anomaly score is higher in the hierarchical structure than the identified group level anomaly score;
- identifying, by the metric processing component, one or more metric attributes that are determined to be a source of the identified group level anomaly score;
- and
- generating, by the metric processing component, a processing result corresponding to the identified system level, component level and group level anomaly scores.
5 more elements
A system to manage the processing of operations information comprising:
- one or more virtual machine instances executing applications, wherein individual virtual machine instances collect operations information that includes at least one metric attribute and transmit the collected operations information;
- and
- a metric processing component implemented on at least one computing device including a processor and a memory, the metric processing component configured to receive the collected operations information and manage the processing of the collected operations information, the metric processing component configured to:
- obtain the collected operations information from the individual virtual machine instances, the collected operations information including a plurality of metric attributes;
- generate a hierarchical structure defining anomaly scores for individual levels of the hierarchical structure based on machine learning, the anomaly scores include one or more group level anomaly scores formed from the obtained one or more metric attributes, one or more component level anomaly scores formed from the one or more group level anomaly scores, and a system level anomaly score formed from the one or more component level anomaly scores;
- identify, using the hierarchical structure, the system level anomaly score indicative of an anomaly, a component level anomaly score that is determined to be a source of the identified system level anomaly score, a group level anomaly score that is determined to be a source of the identified component level anomaly score, and a metric attribute of the one or more metric attributes that is determined to be a source of the anomaly, wherein the identified system level anomaly score is higher in the hierarchical structure than the identified component level anomaly score and wherein the identified component level anomaly score is higher in the hierarchical structure than the identified group level anomaly score;
- and
- generate a processing result based on processing the hierarchical structure and corresponding to determination of the anomaly.
5 more elements
A computer-implemented method to manage operations information comprising:
- obtaining metric attributes from agents;
- generating a hierarchical structure identifying anomaly scores for individual levels of the hierarchical structure based on machine learning, wherein the anomaly scores include one or more group level anomaly scores formed from the obtained metric attributes, one or more component level anomaly scores formed from the one or more group level anomaly scores, and a system level anomaly score formed from the one or more component level anomaly scores;
- processing the hierarchical structure to identify a likely source of an anomaly by identifying the system level anomaly score indicative of the anomaly, a component level anomaly score that is determined to be a source of the identified system level anomaly score, a group level anomaly score that is determined to be a source of the identified component level anomaly score, and the one of the metric attributes that is determined to be a source of the identified group level anomaly score, wherein the identified system level anomaly score is higher in the hierarchical structure than the identified component level anomaly score and wherein the identified component level anomaly score is higher in the hierarchical structure than the identified group level anomaly score;
- and
- determining a processing result associated with the anomaly.
2 more elements
Dependencies are inferred from the claim text.