Rule-based adaptive monitoring of application performance

Patent metadata

Field	Value
Publication number	US-10078571-B2
Application number	US-201514963939-A
Country	US
Kind code	B2
Filing date	Dec 9, 2015
Priority date	Dec 9, 2015
Publication date	Sep 18, 2018
Grant date	Sep 18, 2018

Abstract

A method for dynamically and adaptively monitoring a system based on its running behavior adjusts monitoring levels of the monitored application in real-time. A rules-based mechanism dynamically adjusts monitoring levels in real-time, based on the system's performance observed during a workload run, whether in a production or test environment.

First claim

What is claimed is:

1. A method for adaptive performance monitoring of an enterprise-level distributed system, comprising:
using a processor device in operative communication with a performance test tool, performing:
prior to run time, initializing low default monitoring levels at logging points distributed across a plurality of system nodes comprising multiple hardware and software components in the enterprise-level distributed system handling thousands of concurrent users;
initializing a rules set with a set of rules associated with a plurality of system nodes under evaluation, wherein the rules specify, for each of the plurality of system nodes:
at least one performance goal expressed as a business rule tied to a service level agreement;
associated with the at least one performance goal, a monitoring adjustment to perform in event of a rule failure, the monitoring adjustment defining a required change in monitoring output at the logging points associated with the system node to capture an increased amount of monitoring data, wherein the monitoring adjustment specifies a number of monitoring logging levels to increase; and
a specified period of time to maintain the monitoring adjustment, after which time the monitoring logging levels revert to their low default monitoring levels so as not to deleteriously affect system performance;

storing the rules set in a data store; and
iteratively performing at pre-determined intervals during run-time for each system node under evaluation:
using the performance test tool, collecting monitoring data from the logging points;
analyzing the monitoring data to derive performance metrics;
accessing the rules set;
performing a comparison between the derived performance metrics and the at least one performance goal in the rules set to determine if the rule failure has occurred;
matching the performance goal associated with the rule failure to its associated monitoring adjustment;
automatically performing the associated monitoring adjustment stated in the rules set; and
automatically reverting to the default monitoring levels after the specified period of time has elapsed.

2. The method of claim 1 wherein automatically performing the associated monitoring adjustment comprises:
adjusting the monitoring output for the system node associated with the rule failure.

3. The method of claim 1 further comprising:
adding a filter to the rules set to produce a complex rule incorporating additional calculations depending on severity of the rule failure;
wherein automatically performing the associated monitoring adjustment further comprises:
applying the filter to determine whether the derived performance metrics fail to meet the at least one performance goal by a specified threshold amount; and
further changing the monitoring output to a maximum level during run-time, responsive to the determining.

4. The method of claim 3 wherein further changing the monitoring output comprises at least one of: adjusting sampling size, adjusting workload, adding logging points, and adjusting the pre-determined intervals.

5. The method of claim 1 wherein collecting the monitoring data comprises collecting monitoring data measuring at least one of: cpu utilization, deadlock occurrences, and access response times.

6. The method of claim 1 further comprising updating the rules set based on monitoring results.

7. An information processing system for adaptive performance monitoring of a distributed system, the information processing system comprising:
a memory;
a processor communicatively coupled to the memory and a performance test tool, the processor being configured to perform a method comprising:
initializing low default monitoring levels at logging points prior to run-time;
initializing a rules set with a set of rules associated with a plurality of system nodes under evaluation, wherein the rules specify, for each of the plurality of system nodes:
at least one performance goal expressed as a business rule tied to a service level agreement;
associated with the at least one performance goal, a monitoring adjustment to perform in event of a rule failure, the monitoring adjustment defining a required change in monitoring output at the logging points associated with the system node to capture an increased amount of monitoring data, wherein the monitoring adjustment specifies a number of monitoring logging levels to increase; and
a specified period of time to maintain the monitoring adjustment, after which time the monitoring logging levels revert to their low default monitoring levels so as not to deleteriously affect system performance;

storing the rules set in a data store; and
iteratively performing at pre-determined intervals during run-time for each system node under evaluation:
using the performance test tool, collecting monitoring data from the logging points;
analyzing the monitoring data to derive performance metrics;
accessing rules set;
performing a comparison between the derived performance metrics and the at least one performance goal in the rules set to determine if the rule failure has occurred;
matching the performance goal associated with the rule failure to its associated monitoring adjustment;
automatically performing the associated monitoring adjustment stated in the rules set; and
automatically reverting to the default monitoring levels after the specified period of time has elapsed.

8. The information processing system of claim 7 wherein the monitoring data measures at least one of: cpu utilization, deadlock occurrences, and access response times.

9. The information processing system of claim 7 wherein the monitoring adjustment comprises at least one of: adjusting sampling size, adjusting workload, adding logging points, and adjusting the pre-determined intervals.

10. The information processing system of claim 7 wherein the method for automatically performing the associated monitoring adjustment comprises:
increasing the monitoring output for the system node reporting the rule failure.

11. The information processing system of claim 10 further comprising:
a filter added to the rules set to produce a complex rule incorporating additional calculations depending on severity of the rule failure;
wherein the method for automatically performing the associated monitoring adjustment further comprises:
applying the filter to determine whether the derived performance metrics fail to meet the at least one performance goal by a specified threshold amount; and
further changing the monitoring output to a maximum level during run-time, responsive to the determining.

12. The information processing system of claim 7 wherein the method further comprises updating the rules set based on monitoring results.

13. A computer program product for adaptive performance monitoring of a distributed system, the computer program product comprising:
a non-transitory storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising:
initializing low default monitoring levels at logging points prior to run-time;
initializing a rules set with a set of rules associated with a plurality of system nodes under evaluation, wherein the rules specify, for each of the plurality of system nodes:
at least one performance goal expressed as a business rule tied to a service level agreement;
associated with the at least one performance goal, a monitoring adjustment to perform in event of a rule failure, the monitoring adjustment defining a required change in monitoring output at logging points associated with the system node to

Assignees

• IBM
• Univ Dublin

Inventors

• Altman Erik R
• Assem Aly Salama Hitham Ahmed
• Mitchell Nicholas M
• O'Sullivan Patrick Joseph
• Portillo Dominguez Andres Omar
• Sweeney Peter F

Classifications

• G06F11/076

  by exceeding a count or rate limit, e.g. word- or bit count limit · CPC title

• G06F11/302

  where the computing system component is a software system · CPC title

• G06F11/0706

  the processing taking place on a specific hardware platform or in a specific software environment · CPC title

• G06F11/079

  Root cause analysis, i.e. error or fault diagnosis (in a hardware test environment G06F11/22; in a software test environment G06F11/36) · CPC title

• G06F11/3495 Primary

  for systems · CPC title