Performance Bottleneck Identification Through Software Diagnostics- Impetus Webinar

8/4/2019 Performance Bottleneck Identification Through Software Diagnostics- Impetus Webinar

1/18

Identifying Software Performance

Bottlenecks using Diagnostic Tools

by

Impetus Technologies

Recorded version available at

http://www.impetus.com/webinar_registration?event=archived&eid=46


2/18

Outline

Typical process of bottleneck identification

Diagnostic tools

How diagnostic tools work?

Diagnostic tools usage scenarios

Server side diagnostics

Client side diagnostics

Benefits of diagnostic tools

Real world examples

Q&A




3/18

Typical Process of Bottleneck Identification

Extensive time, resources, and efforts involved


4/18

Diagnostic Tools

Response time break down

Tiers, components and layers

Information

Bottleneck methods, costliest queries, memory leaks

Deployment options

Development environment, load testing environment,

production

Differ from profilers

Low overheads, isolates code bottlenecks




5/18

How Diagnostic Tools Work?


6/18

Server-side Diagnostics: Plug-in Tools

An extension to performance testing tools

Diagnostic data integrates with test results

Break down of transaction response time

Diagnostic configuration a part of performance testing tool

Easy analysis

Some popular tools available- Dynatrace, Deep Diagnostics




7/18

Server-side Diagnostics: Custom Tools

Not integrated with any performance testing tool

Runs in addition to the performance tests

Independent configuration and analysis

Existing tools, scripts can be used

Support for specific technology, requirements

Some popular tools available- Glassbox, InfraRed




8/18

Server-side Diagnostics: Approach

Select appropriate diagnostic tool Objective

Technology

Cost

Calibrate overhead of diagnostic tool Apply method filters

Bottleneck identification

Top X methods and queries

Response time distribution across layers




9/18

Client-side Diagnostics

Vastly impacts end user experience

Categories of client side

Mobile diagnostics

Measuring app performance on mobile device

Limited hardware resources

rowser diagnostics

Identifies browser bottlenecks in page loading




10/18

Client-side Diagnostics: Approach

Mobile diagnostics approach

Existing tools (J2ME profiler, Android Trace)

Custom solution

Use of Aspect oriented frameworks (AspectME, AspectLUA)

Use byte code instrumentation

Re-usable aspects

Configure aspect within application binaries

rowser diagnostics approach

Diagnostics plugins - Firebug, SpeedTrace, HttpWatch

Analyze statistics

Identify bottlenecks - Java script execution, memory leaks, page

loading




11/18

Benefits of Diagnostic Tools

Drill down user transactions to methods, DB queries and web

service calls

Performance visibility across complete application

Identifies and isolates performance bottlenecks

Costly queries, high method invocations, method execution time,suspicious memory consumption

Enhanced collaboration

Rich analysis dashboard

Built-in heuristics Integrates with IDE

Reduces time and efforts involved in bottleneck identification


12/18

The Client One of the leading companies in home health care

Key Challenge

Performance bottleneck identification from performance testing

Lack of expertise in diagnostic tools

Impetus Solution

Proposed SandStorm and its diagnostic plug-in

Calibrated diagnostics overhead (~2%-3%)

Captures diagnostics data across method & queries during test execution

Performed client side diagnostics using Firebug

Consolidated data enabled quick isolation of performance bottlenecks

Case Study: Plug-in Diagnostic Tools

enefits Realised

Faster performance bottleneck identification

Streamlined and efficient performance engineering

No performance issues in production


13/18

Case Study: Custom Diagnostic Tools

The Client Leading provider of financial risk management solutions

Key Challenge

Mechanism to isolate code bottlenecks during performance testing

Performance monitoring of application in staging environment

Reusing Load Runner scripts & performance center scenarios

Impetus Solution

Proposed implementing an extensible, custom diagnostic solution

Built diagnostic f/w using Aspect Oriented Programming using

AspectWerkz, Jamon, Jarep libraries

Deployed and configured utility in a staging environment

Reports generated & analysed to identify expensive queries

enefits Realised

Monitoring mechanisms in staging environment

Cost saving as same performance testing tool was used


14/18

Application Diagnostics Utility

Server side diagnostic plug-in to SandStorm

Drill down transaction response times across multiple tiers andcomponents

Identifies slow performing methods and queries

Easy configuration

No code changes required

User selection of monitored methods

Low overheads

Supports multiple Java app servers and programming stacks




15/18

Summing Up

Diagnostic tools differ from profilers

Diagnostic tools can be used in multiple environments

Software diagnostics an integral piece of a successful

performance engineering strategy

Custom diagnostic solutions for specific requirements

Calibrate the overhead of diagnostic tool before actual usage




16/18

About Us

Strategic partners for software product engineering and R&D

Thought leaders in cutting-edge technologies

Mature processes and practices that are methodical, yet flexible

Diverse domain expertise

Holistic approach in Performance Engineering




17/18

Questions

Please send in your questions

using the chat panel




18/18

Thank youFor more information,

write to us at [email protected]

@impetuscalling



Performance Bottleneck Identification Through Software Diagnostics- Impetus Webinar

Documents