Root cause analysis of server system performance degradations
First Claim
1. A computer-implemented method of facilitating analysis of the operation of a transactional server system, the method comprising:
- analyzing a series of measurement values of a first performance parameter to detect a degradation in performance of the transactional server system as seen by end users, said series of measurement values representing measurements taken over a period of time;
identifying a plurality of additional parameters for which a predefined dependency relationship exists with the first performance parameter, wherein each of the additional parameters is associated with a particular potential source of the performance degradation, and is lower in level than the first performance parameter;
analyzing measurement values of each of the plurality of additional parameters to evaluate a likelihood that each such additional parameter is associated with the performance degradation, said measurement values of the additional parameters representing measurements taken over said period of time; and
presenting to a user an indication of which of the additional parameters is likely associated with the performance degradation to assist the user in identifying a root cause of the performance degradation.
10 Assignments
0 Petitions
Accused Products
Abstract
A system for monitoring and analyzing the post-deployment performance of a web-based or other transactional server is disclosed. The monitoring system includes agent components that monitor and report various performance parameters associated with the transactional server, such as response times seen by end users, server and network times, and various server resource utilization parameters. A web-based reports server displays the data collected by the agents through a series of charts and graphs that indicate whether correlations exist between the response times and lower level parameters. A root cause analysis system applies statistical algorithms to the collected data to detect performance degradations in specific parameters, and uses predefined parameter dependency rules to correlate high level performance problems to likely sources or causes of such problems.
393 Citations
24 Claims
-
1. A computer-implemented method of facilitating analysis of the operation of a transactional server system, the method comprising:
-
analyzing a series of measurement values of a first performance parameter to detect a degradation in performance of the transactional server system as seen by end users, said series of measurement values representing measurements taken over a period of time;
identifying a plurality of additional parameters for which a predefined dependency relationship exists with the first performance parameter, wherein each of the additional parameters is associated with a particular potential source of the performance degradation, and is lower in level than the first performance parameter;
analyzing measurement values of each of the plurality of additional parameters to evaluate a likelihood that each such additional parameter is associated with the performance degradation, said measurement values of the additional parameters representing measurements taken over said period of time; and
presenting to a user an indication of which of the additional parameters is likely associated with the performance degradation to assist the user in identifying a root cause of the performance degradation. - View Dependent Claims (2, 3, 4, 5, 6, 7)
-
-
8. A computer-implemented method of evaluating root causes of performance degradations in a transactional server, comprising:
-
monitoring an end user performance parameter reflective of the performance of the transactional server as seen by end users;
monitoring a plurality of server resource parameters of the transactional server;
analyzing data values of the end user performance parameter to detect a degradation in end user performance; and
analyzing data values of each of the plurality of server resource parameters to evaluate a likelihood that each such server resource parameter is associated with a cause of the degradation in end user performance;
whereby server resource variations that cause degradations in end user performance are detected automatically. - View Dependent Claims (9)
-
-
10. A system for analyzing the performance of a server system, comprising:
-
a data repository which stores performance data collected by a set of agents that monitor operation of the server system, the performance data including measurements of end-user performance parameters indicative of performance experienced by end users of the server system, and including measurements of source parameters indicative of potential sources of degradations in end user performance, said measurements of end-user performance parameters and source parameters taken over a common period of time; and
an application component that analyzes the performance data to identify correlations between degradations in end user performance and degradations in specific source parameters, to thereby predict likely causes of the degradations in end user performance. - View Dependent Claims (11, 12, 13, 14)
-
-
15. A computer program for identifying and predicting root causes of performance problems associated with a server system, comprising, on a computer-readable medium:
-
a first component that analyzes performance data for the server system to identify performance problems associated with the server system, the first component generating a data structure representing performance dependencies between various metric measurements for metrics associated with performance of the server system; and
a second component that displays a user interface in which metric measurements represented in the data structure are grouped into nodes of a hierarchical node structure displayed to a user, the second component displaying in the hierarchical node structure an indication of whether the metric measurements grouped in each individual node represent a performance degradation relative to historical measurements. - View Dependent Claims (16)
-
-
17. A computer-implemented method of identifying likely root causes of performance degradations in a server system, the method comprising:
-
monitoring a response time of the server system over a period of time to generate a series of response time values;
monitoring a plurality of server resource parameters of the server system over said period of time to generate a series of values of each server resource parameter, wherein each of the server resource parameters is associated with a potential cause of degradations in the response time of the server system; and
analyzing the series of response time values and each series of server resource parameter values to evaluate a degree to which degradations in quality of specific server resource parameters are correlated with degradations in the response time. - View Dependent Claims (18, 19, 20, 21, 22, 23, 24)
-
Specification