Detecting and remediating root causes of performance issues

Patent metadata

Field	Value
Publication number	US-10554514-B2
Application number	US-201615335310-A
Country	US
Kind code	B2
Filing date	Oct 26, 2016
Priority date	Oct 26, 2016
Publication date	Feb 4, 2020
Grant date	Feb 4, 2020

Abstract

Exemplary methods, apparatuses, and systems include receiving time series data for each of a plurality of performance metrics. The time series data is sorted into buckets based upon an amount of variation of time series data values for each performance metric. The time series data in each bucket is divided into first and second clusters of time series data points. The bucket having the greatest distance between clusters is used to determine a performance metric having a greatest distance between clusters. The performance metric having the greatest distance between clusters is reported as a potential root cause of a performance issue.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
receiving time series data values for each of a plurality of performance metrics for a computing environment;
determining an amount of variation of time series data values for each performance metric within the plurality of performance metrics;
in response to determining the amount of variation, assigning, for each performance metric, the performance metric time series data values to a bucket, of a plurality of buckets, based upon the amount of variation, wherein each bucket has a particular range of values for the amount of variation;
clustering time series data values assigned to one or more buckets by dividing time series data values for the performance metrics in each bucket into first and second clusters of time series data values;
determining which of the buckets has been divided into first and second clusters with a greatest distance between clusters;
determining, within the bucket having the greatest distance between clusters, a performance metric having a greatest distance between clusters;
recursively bucketing and clustering time series data values within each of the first and second clusters of the bucket determined to have the greatest distance between clusters until the time series data values within a set of buckets cannot successfully be divided into first and second clusters; and
generating a report identifying the performance metric having the greatest distance between clusters within one or more levels of recursion as a potential root cause of a performance issue.

2. The computer-implemented method of claim 1 , wherein the amount of variation of time series data values for each performance metric is a standard deviation value for the time series data values for that performance metric.

3. The computer-implemented method of claim 1 , further comprising for each of one or more levels of recursion:
determining which of the buckets within the level of recursion has been divided into first and second clusters with a greatest distance between clusters;
determining, within the bucket having the greatest distance between clusters within the level of recursion, a performance metric having a greatest distance between clusters; and
generating a report identifying the performance metric having the greatest distance between clusters within the level of recursion as another potential root cause of a performance issue.

4. The computer-implemented method of claim 1 , wherein a bucket cannot successfully be divided into first and second clusters when the time series data values within the bucket lacks a threshold amount of difference.

5. The computer-implemented method of claim 1 , wherein a bucket cannot successfully be divided into first and second clusters when the performance metric having the greatest distance between clusters in a parent level of recursion would also be the performance metric having the greatest distance between clusters in a current level of recursion.

6. The computer-implemented method of claim 1 , wherein, for determining which of the buckets has been divided into first and second clusters with the greatest distance between clusters, distance between clusters is a multidimensional distance.

7. The computer-implemented method of claim 1 , wherein each bucket represents a range of variation of time series data values and the performance metric time series data values are assigned to the bucket in response to determining the amount of variation of time series data values for the performance metric are within the range of variation for the bucket.

8. A non-transitory computer-readable medium storing instructions, which when executed by a processing device, cause the processing device to perform a method comprising:
receiving time series data values for each of a plurality of performance metrics for a computing environment;
determining an amount of variation of time series data values for each performance metric within the plurality of performance metrics;
in response to determining the amount of variation, assigning, for each performance metric, the performance metric time series data values to a bucket, of a plurality of buckets, based upon the amount of variation, wherein each bucket has a particular range of values for the amount of variation;
clustering time series data values assigned to one or more buckets by dividing time series data values for the performance metrics in each bucket into first and second clusters of time series data values;
determining which of the buckets has been divided into first and second clusters with a greatest distance between clusters;
determining, within the bucket having the greatest distance between clusters, a performance metric having a greatest distance between clusters;
recursively bucketing and clustering time series data values within each of the first and second clusters of the bucket determined to have the greatest distance between clusters until the time series data values within a set of buckets cannot successfully be divided into first and second clusters; and
generating a report identifying the performance metric having the greatest distance between clusters within one or more levels of recursion as a potential root cause of a performance issue.

9. The non-transitory computer-readable medium of claim 8 , wherein the amount of variation of time series data values for each performance metric is a standard deviation value for the time series data values for that performance metric.

10. The non-transitory computer-readable medium of claim 8 , the method further comprising for each of one or more levels of recursion:
determining which of the buckets within the level of recursion has been divided into first and second clusters with a greatest distance between clusters;
determining, within the bucket having the greatest distance between clusters within the level of recursion, a performance metric having a greatest distance between clusters; and
generating a report identifying the performance metric having the greatest distance between clusters within the level of recursion as another potential root cause of a performance issue.

11. The non-transitory computer-readable medium of claim 8 , wherein a bucket cannot successfully be divided into first and second clusters when the time series data values within the bucket lacks a threshold amount of difference.

12. The non-transitory computer-readable medium of claim 8 , wherein a bucket cannot successfully be divided into first and second clusters when the performance metric having the greatest distance between clusters in a parent level of recursion would also be the performance metric having the greatest distance between clusters in a current level of recursion.

13. The non-transitory computer-readable medium of claim 8 , wherein, for determining which of the buckets has been divided into first and second clusters with the greatest distance between clusters, distance between clusters is a multidimensional distance.

14. The non-transitory computer-readable medium of claim 8 , wherein each bucket represents a range of variation of time series data values and the performance metric time series data values are assigned to the bucket in response to determining the amount of variation of time series data values for the performance metric are within the range of variation for the bucket.

15. An apparatus comprising:
a processing device; and
a memory coupled to the processing device, the memory storing instructions which, when executed by the processing device, cause the apparatus to:
receive time series data values for each of a

Assignees

• Vmware Inc

Inventors

• Chen Chien-Chia
• Singaravelu Lenin
• Zhou Ruijin
• Huang Xiaobo

Classifications

• H04L41/142

  using statistical or mathematical methods · CPC title

• H04L43/04 Primary

  Processing captured monitoring data, e.g. for logfile generation · CPC title

• H04L43/08

  Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters · CPC title