Related metrics-based monitoring of computing-instances

Patent metadata

Field	Value
Publication number	US-12056039-B2
Application number	US-202117561767-A
Country	US
Kind code	B2
Filing date	Dec 24, 2021
Priority date	Dec 24, 2021
Publication date	Aug 6, 2024
Grant date	Aug 6, 2024

Abstract

In one example, a computing node includes a metric dependency graph knowledge base to store a data structure representing a relationship between a plurality of metrics. Further, the computing node may include a processor and a memory having a metric recommendation unit. The metric recommendation unit may determine a first metric of a monitored computing-instance while a user interacts with a GUI of a monitoring application. Further, the metric recommendation unit may retrieve the data structure corresponding to the first metric. The data structure may include the first metric and a plurality of dependent metrics associated with the first metric. Further, the metric recommendation unit may apply a machine learning model on the data structure to determine a second metric from the plurality of dependent metrics. Furthermore, the metric recommendation unit may output the second metric related to the first metric on the GUI.

First claim

What is claimed is:

1. A computing node comprising:
a metric dependency graph knowledge base to store a data structure representing a relationship between a plurality of metrics;
a processor; and
a memory comprising a metric recommendation unit to:
determine a first metric of a monitored computing-instance while a user interacts with a graphical user interface (GUI) of a monitoring application;
retrieve the data structure corresponding to the first metric from the metric dependency graph knowledge base, the data structure comprising the first metric and a plurality of dependent metrics associated with the first metric;
apply a machine learning model on the data structure to determine a second metric from the plurality of dependent metrics, wherein the machine learning model is trained to determine the second metric related to the first metric based on navigation pattern data of users interacting with the GUI, the machine learning model being trained with the navigation pattern data and the data structure, the data structure and the machine learning model being used to recommend, a set of related metrics for a first metric when a user selects the first metric while interacting with the GUI; and
output the second metric related to the first metric on the GUI.

2. The computing node of claim 1 , wherein the metric recommendation unit is to output the second metric by generating a dashboard including a chart on the GUI, the chart representing the second metric.

3. The computing node of claim 1 , wherein the metric recommendation unit is to enable the user to navigate through the second metric to identify a root cause of an issue associated with the monitored computing-instance.

4. The computing node of claim 1 , wherein the metric recommendation unit is to output the second metric related to the first metric on the GUI in response to detecting an event that occurs in the monitored computing-instance of a data center.

5. The computing node of claim 1 , wherein the data structure is formatted in accordance with one or more of JavaScript object notation (JSON), extensible markup language (XML), a binary file, a database file, YAML ain't markup language (YAML), and/or a proprietary encoding scheme.

6. The computing node of claim 1 , wherein the navigation pattern data comprises time series data captured while the users navigate through the plurality of metrics in the GUI associated with the monitoring application.

7. The computing node of claim 1 , wherein the metric recommendation unit is to:
apply the machine learning model on the data structure to determine a third metric related to the second metric from the plurality of dependent metrics in response to a user selection of the second metric on the GUI; and
output the third metric related to the second metric on the GUI.

8. A non-transitory computer-readable storage medium comprising instructions that, when executed by a processor of a computing node, cause the processor to:
receive a selection of a first metric of a monitored computing-instance while a user interacts with a graphical user interface (GUI) of a monitoring application;
retrieve a data structure corresponding to the first metric from a metric dependency graph knowledge base, the data structure comprising the first metric and a plurality of dependent metrics associated with the first metric;
apply a machine learning model on the data structure to determine a set of related metrics from the plurality of dependent metrics, wherein the machine learning model is trained to determine the set of related metrics for the first metrics based on navigation pattern data of users interacting with the GUI, the machine learning model being trained with the navigation pattern data and the data structure, the data structure and the machine learning model being used to recommend, a set of related metrics for a first metric when a user selects the first metric; and
output the set of related metrics on the GUI.

9. The non-transitory computer-readable storage medium of claim 8 , further comprising instructions to:
create a dashboard including at least one chart on the GUI, the at least one chart representing the set of related metrics to monitor a health of the monitored computing-instance.

10. The non-transitory computer-readable storage medium of claim 8 , further comprising instructions to:
enable the user to navigate through the set of related metrics to identify a root cause of an issue associated with the monitored computing-instance.

11. The non-transitory computer-readable storage medium of claim 8 , wherein the set of related metrics related to the first metric are outputted on the GUI in response to detecting an event that occurs in the monitored computing-instance of a data center.

12. The non-transitory computer-readable storage medium of claim 8 , wherein the data structure is formatted in accordance with one or more of JavaScript object notation (JSON), extensible markup language (XML), a binary file, a database file, YAML ain't markup language (YAML), and/or a proprietary encoding scheme.

13. The non-transitory computer-readable storage medium of claim 8 , wherein the navigation pattern data comprises time series data captured while the user navigates through various charts in the GUI associated with the monitoring application, and wherein a plurality of metrics is displayed in different types of charts.

14. The non-transitory computer-readable storage medium of claim 8 , wherein instructions to apply the machine learning model on the data structure to determine the set related metrics related to the first metric comprise instructions to:
filter the plurality of dependent metrics by applying the machine learning model to the selected first metric.

15. A computer-implemented method comprising:
receiving metrics and relationship between the metrics associated with a monitored computing instance running in a data center;
generating a data structure including metric dependency levels associated with the metrics based on the relationship between the metrics;
storing the data structure in a metric dependency graph knowledge base;
receiving navigation pattern data of users interacting with a graphical user interface (GUI) of a monitoring application that monitors the monitored computing instance, wherein the navigation pattern data is captured when the users browse through the metrics in the GUI;
building a machine-learning model to determine related metrics for each given metric by training the machine-learning model with the navigation pattern data and the data structure; and
utilizing the data structure and the machine learning model to recommend, in real time, a set of related metrics for a first metric when a user selects the first metric while interacting with the GUI.

16. The computer-implemented method of claim 15 , wherein the navigation pattern data comprises a plurality of screenshots captured during a sequence of user interaction with various metrics of the GUI over a period.

17. The computer-implemented method of claim 15 , wherein building the machine-learning model comprises:
training a set of machine learning models to predict related metrics for each given metric using a train dataset of the navigation pattern data;
validating the trained machine learning models to tune an accuracy of the trained machine learning models based on a validation dataset of the navigation pattern data;
testing the trained set of machine learning models with a test dataset of the navigation pattern data; and
determining the machine learning model having a maxi

Assignees

• VMware LLC

Inventors

• Govindaraju Agila
• Dhamale Rutuja
• Gurram Charitha
• Bangera Ramya
• Rout Auroprava

Classifications

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06F9/542

  Event management; Broadcasting; Multicasting; Notifications · CPC title

• G06F9/451

  Execution arrangements for user interfaces · CPC title

• G06F2009/45591

  Monitoring or debugging support · CPC title

• G06N20/00

  Machine learning · CPC title