13
Luba Ilieva, Senior QA Engineer Musala Soft Ltd. 15.February.2007 Automating Performance Testing with JMeter

Automating Performance Testing JMeter Luba Ilieva

Embed Size (px)

Citation preview

Page 1: Automating Performance Testing JMeter Luba Ilieva

Luba Ilieva, Senior QA EngineerMusala Soft Ltd.

15.February.2007

Automating

Performance

Testing with JMeter

Page 2: Automating Performance Testing JMeter Luba Ilieva

2

Agenda

• General Truths about Performance

• Performance Test Design Approaches

• JMeter Architecture

• Practical Solution with JMeter

• Demo

Page 3: Automating Performance Testing JMeter Luba Ilieva

3

General Truths about Performance

• Doubling an attribute, like CPU speed, doesn’t necessarily increase performance 2X

• Increasing a resource that is not bottleneck will not necessarily increase performance

• With respect to cache, size matters

Page 4: Automating Performance Testing JMeter Luba Ilieva

4

Test Design Approaches

• A brand new application– Table of comparison between environments

• Application is on production– Comparison analysis between versions

Page 5: Automating Performance Testing JMeter Luba Ilieva

5

The Test Tool – Apache JMeter

• Apache JMeter is100% pure Java desktop application designed to load test functional behavior and measure performance. It was originally designed for testing Web Applications but has since expanded to other test functions

• Java based

• Highly extensible

• URL: http://jakarta.apache.org/jmeter/

Page 6: Automating Performance Testing JMeter Luba Ilieva

6

JMeter Architecture

Test Plan:

– Thread Groups

– Controllers

– Samplers

– Listeners

– Config elements

– Assertions

– Timers

– Other Elements

Page 7: Automating Performance Testing JMeter Luba Ilieva

7

JMeter Outlook

Selected Element

Active Element

Deactivated Element

Properties Area

Run Test Indicator

Page 8: Automating Performance Testing JMeter Luba Ilieva

8

Practical Case - Background Info

• Technology: webMethods IS, Java, Oracle 9i

• Goals: Identify bottlenecks in webMethods components

• Approach: Simulated production peak hours load

Call via HTTP

Standard System Inbox

SUT

Integration

Server

DB

JBoss

AS

JMeter

(Test Driver)

App. specific

extension

Page 9: Automating Performance Testing JMeter Luba Ilieva

9

Analyzing Results –JMeter Metrics

JMeter Tree View

• Node for every request

• Properties of the request

Request Properties Area

Failed request

Successful Request

Page 10: Automating Performance Testing JMeter Luba Ilieva

10

Analyzing Results –Server Utilization

Windows Performance Monitor Utility

Perfmon.exe

• CPU utilization• VM utilization• Thread Count

Process Virtual memory

CPU utilizatio

n

Thread Count

Page 11: Automating Performance Testing JMeter Luba Ilieva

11

Analyzing Results –Manual Correction

• Runtime

– Monitor PerfMon GUI and JMeter Tree View Results

• Recorded data– JMeter CSV file (timestamp, request time, request

name, response status)

– Excel Spreadsheet

Page 12: Automating Performance Testing JMeter Luba Ilieva

JMeter

Demonstration

Page 13: Automating Performance Testing JMeter Luba Ilieva

Thank You!

e-mail: [email protected]

presentation:http://www.musala.com/ppt/MusalaSoft_Perf_Testing_JMeter.pdf