PERFORMANCE TEST
EmbedIT
21st April 2015
Petr Střítecký, Michal Godar
Definition
<2>
˂ Performance test determines how a system performs in items of
responsiveness and stability under a particular workload
˂ It can also serve to investigate, measure, validate or verify other
quality attributes of the system, such as scalability, reliability and
resource usage
http://en.wikipedia.org/wiki/Software_performance_testing
˂ Prove that the system in a new version is stable under the real
workload
˂ Prevent future performance issues
˂ Perform tests which can‘t be executed manually
˂ Help to increase system overall stability
˂ Set the limits and find the bottlenecks of the application
˂ Ensure that application will be ready for business attractive terms
Motivation
<3>
Processes architecture and technologies
<4>
Processes architecture and technologies
<5>
Processes architecture and technologies
<6>
Processes architecture and technologies
˂ Performance test prerequisites
˂ non-functional requirements
˂ functional specification
˂ Test data preparation
˂ static data
˂ random generated data
˂ load required data from database
˂ production data scaling - synthetic data
˂ Issue reporting
˂ test run evaluation
˂ issues are created in tracking system
<7>
Processes architecture and technologies
˂ Silk Performer
˂ Overall management of the performance test
Application recording
Code adjustments
Workload and agents configuration
Test run and monitoring
Results analysis
˂ Structural programming language
˂ Wide protocol support
˂ Front-end diagnostics from the end user perspective - TrueLog
<8>
Processes architecture and technologies
<9>
˂ Other supported protocols
˂ Expected scenarios
˂ Non-functional requirematns
<10>
Baseline test
Requirements
˂ Workload
Questionnaire Workload
Baseline test
Run
<11>
Monitor
Baseline test
HW Monitoring
˂ Database
˂ Oracle Enterprise Manager
˂ AWR Report
˂ ASH Report
˂ Nagios
<12>
˂ Application server
˂ WebLogic DomainHealth
˂ Application logs
˂ Oracle JRockit
˂ Nagios
Baseline test
Results
˂ Analyze client site data
<13>
˂ Report
Other test
<14>
˂ Stress test
˂ Finding application limits and bottlenecks
˂ Module test
˂ Detailed, focused on internal module performance
˂ What-if test
˂ Simulate unusual state and check if the system is able to handle it
˂ Comparative HW test
˂ Same application is tested on different HW platform
Questions?
Conclusion
<15>