From time to time, you may notice variances within the Loadrunner Analysis tool. This is most apparent when reviewing the Transaction Response Time (Percentile) graph.
If you look at the following Transaction Summary Report:
And look at the following Transaction Response Time (Percentile) graph:
As you can see there are some conflicting values:
Summary Report | Graph Legend | Graph Mouseover (Tooltip) | ||
Average | 50th Percentile | Average | Median | 50th Percentile |
7.718 | 5.728 | 7.942 | 5.830 | 5.728 |
Firstly, I'll state that all these values are correct. The LoadRunner Analysis engine calculates all values correctly, and all variances can be easily explained by examining the data.
Median vs Mean vs Average vs Percentile
The first concept to understand and permanently implant into your brain, is the difference between these four aggregate types. An "Aggregate" is a single value that represents an underlying set of multiple values, commonly known aggregates are Average, Minimum and Maximum.
Most people understand what an "Average" is...that is the sum of a value set divided by the count of the value set. "Average" is also known as "Mean". "Median" however is a different concept, a median value is calculated by sorting a set of data values from smallest to largest value then dividing this sorted set by two. The "Median" is the middle value at the place you divided the set. Median is identical to the 50th Percentile. Percentiles are an extension on the median concept, but instead of dividing the sorted value set by two...we instead divide the value set by a percentage (starting with the smallest value). So the 15th Percentile, refers to the value 15% along the sorted set...the 75th Percentile refers to the value 75% along the sorted set. Also the 0th Percentile is undefinable value.
Based on this information, we can safely separate the values into the following two groups.
Average | Median/ 95th Percentile | |
Summary Report | 7.718 | 5.728 |
Graph Legend | 7.942 | 5.830 |
Graph Mouseover | - | 5.728 |
Clearly the median value for the Graph Mouseover and the Summary Report are identical, so for the rest of the article I'll will no longer refer to the Graph Mouseover.
Graph Values vs Raw Values
When the Percentile graph is generated (or any other graph for that matter), the analysis tool generates a small sample set for the graph. The sample set size is configurable in some graphs by setting the graph's granularity, however the Percentile graph is fixed at a sample set of 100 data values...with each corresponding to a separate percentile value.
Therefore the values for Average and Median for the Graph Legend are calculated from the Graph data, not the Raw data.
Keeping this in mind, the Graph Average value is simply the average of the 1th, 2nd, 3rd...100th Percentile. Whilst this is graph average is close, it is not as accurate as the Summary Report average (which is calculated from the complete raw data set).
Okay, so up to this point we've explained all the value variations, except for the difference with the Median value. This problem is interesting, remember I mentioned that the median value is simply the sorted value set divided into two...and that the 0th Percentile is undefined. Well, the analysis tool equally divides the graph data result set into a set of values representing the 1th Percentile through to the 50th Percentile, and the 51st Percentile through to the 100th Percentile. As there is no clearly defined midpoint in this data set the analysis engine uses the next value...in this case the 51st Percentile.
So in summary. the Graph Median value actually the value of the 51st Percentile...not the 50th Percentile. This may be considered a minor bug.
In Summary
In summary, the graph value data is adequate for producing a graph...however for aggregate values I recommend only using the Summary Report data as the Summary report is the only report in the analysis engine that is guaranteed to display accurate data based on the complete data set.
Tech tips from JDS

Browser Console
Read More

Glide Variables
Read More

Understanding Database Indexes in ServiceNow
Read More

Fast-track ServiceNow upgrades with Automated Testing Framework (ATF)
Read More

Read More

Splunk .conf18
Read More

ServiceNow Catalog Client Scripts: G_Form Clear Values
Read More

Is DevPerfOps a thing?
Read More

The benefits of performance testing with LoadRunner
Read More

Monitoring Atlassian Suite with AppDynamics
Read More

5 quick tips for customising your SAP data in Splunk
Read More

How to maintain versatility throughout your SAP lifecycle
Read More

How to revitalise your performance testing in SAP
Read More

Reserve and import data through Micro Focus ALM
Read More

How to effectively manage your CMDB in ServiceNow
Read More

ServiceNow and single sign-on
Read More

How to customise the ServiceNow Service Portal
Read More

Integrating a hand-signed signature to an Incident Form in ServiceNow
Read More

Integrating OMi (Operations Manager i) with ServiceNow
Read More

Implementing an electronic signature in ALM
Read More

Service portal simplicity
Read More

Learning from real-world cloud security crises
Read More

Static Variables and Pointers in ServiceNow
Read More

Citrix and web client engagement on an Enterprise system
Read More

Understanding outbound web services in ServiceNow
Read More

How to solve SSL 3 recording issues in HPE VuGen
Read More

How to record Angular JS Single Page Applications (SPA)
Read More

Calculating Pacing for Performance Tests
Read More

Vugen and GitHub Integration
Read More

What’s new in LoadRunner 12.53
Read More

Filtered Reference Fields in ServiceNow
Read More

ServiceNow performance testing tips
Read More

Monitor Dell Foglight Topology Churn with Splunk
Read More

Straight-Through Processing with ServiceNow
Read More

Splunk: Using Regex to Simplify Your Data
Read More

ServiceNow Choice List Dependencies
Read More

Tips for replaying RDP VuGen scripts in BSM or LoadRunner
Read More

Incorporating iSPI metric reports into MyBSM dashboard pages
Read More

Using SV contexts to simulate stored data
Read More

What’s new in LoadRunner 12.02
Read More

Recycle Bin for Quality Center
Read More

LoadRunner Correlation with web_reg_save_param_regexp
Read More

LoadRunner 11.52
Read More

QC for Testers – Quiz
Read More

Agile Performance Tuning with HP Diagnostics
Read More

What’s new in HP Service Virtualization 2.30
Read More

Understanding LoadRunner Virtual User Days (VUDs)
Read More

Problems recording HTTPS with VuGen
Read More

Improving the management and efficiency of QTP execution
Read More

Performance testing Oracle WebCenter with LoadRunner
Read More

Generating custom reports with Quality Center OTA using Python
Read More

Asynchronous Communication: Scripting For Cognos
Read More

How to fix common VuGen recording problems
Read More

Monitoring Active Directory accounts with HP BAC
Read More

URL Attachments in Quality Center
Read More

What’s new in LoadRunner 11.00?
Read More

Restore old License Usage stats after upgrading Quality Center
Read More

Changing LoadRunner/VuGen log options at runtime
Read More

Restricting large attachments in Quality Center
Read More

Retrieving Quality Center user login statistics
Read More

A comparison of open source load testing tools
...
Read More

Worst practices in performance testing
Read More

LoadRunner Sales Questions
Read More

LoadRunner Analysis: Hints and tips
Read More

LoadRunner in Windows 7
HP Loadrunner 11 is now available. This new version now natively supports Windows 7 and Windows Server 2008. I ...
Read More

Using the QuickTest Professional “commuter” license
Read More

Installing HP Diagnostics
Read More

Understanding LoadRunner licensing
Read More

VuGen scripting for YouTube video
Read More

Creating a Web + MMS vuser
Read More

Why you should use backwards dates
Read More

How to get the host’s IP address from within VuGen
Read More

VuGen scripting for BMC Remedy Action Request System 7.1
Read More

Unique usernames for BPM scripts
Read More

Mapping drives for LoadRunner Windows monitoring
Read More

VuGen feature requests
Read More

LoadRunner script completion checklist
Read More

Querying Quality Center user roles
Read More

Querying the Quality Center Database
Read More

HPSU 2009 Presentation – Performance Testing Web 2.0
Read More

Scaling HP Diagnostics
Read More

Global variables aren’t really global in LoadRunner
Read More

Client-side certificates for VuGen
Read More

Detect malicious HTML/JavaScript payloads with WebInspect (e.g. ASPROX, Gumblar, Income Iframe)
Read More

VuGen code snippets
Read More

Integrating QTP with Terminal Emulators
Read More

Why you must add try/catch blocks to Java-based BPM scripts
Read More

Querying a MySQL database with LoadRunner
Read More

ANZTB 2009 Presentation: Performance Testing Web 2.0
Read More

How to make QTP “analog mode” steps more reliable
Read More

Testing multiple browsers in a Standardized Operating Environment (SOE)
Read More

DNS-based load balancing for virtual users
Read More

What’s new in LoadRunner 9.50?
Read More

Calculating the difference between two dates or timestamps
Read More

The “is it done yet” loop
Read More

Think time that cannot be ignored
Read More

Understanding aggregate variance within LoadRunner analysis
Read More

Load balancing vusers without a load balancer
Read More

Harvesting file names with VuGen
Read More

Parameterising Unix/Posix timestamps in VuGen
Read More

HP Software trial license periods
Read More

How to handle HTTP POSTs with a changing number of name-value pairs
Read More

VuGen string comparison behaviour
Read More

Persistent data in VuGen with MySQL
Read More

How to write a Performance Test Plan
Read More

Unable to add virtual machine
To get ...
Read More

LoadRunner scripting languages
Read More

WDiff replacement for VuGen
Read More

Testing web services with a standard Web Vuser
Read More

Why your BPM scripts should use Download Filters
Read More

Querying your web server logs
Read More

Importing IIS Logs into SQL Server
Read More

QTP “Uninstall was not completed” problem
Read More

VuGen correlation for SAP Web Dynpro
Read More

How to save $500 on your HP software license
Read More

Testing and monitoring acronyms
Read More

Solving VuGen script generation errors
Read More

An introduction to SiteScope EMS Topology
Read More

Using the BAC JMX Console
Read More
Which response time we should consider?
Average response time or 90 percentile response time
Thanks,
Rudra
Hi,
What you have explained in above section was so good .. here I have a doubt , Like percentile graph average transaction response time also take a set of values to generate graph ?
Thanks
N V Narasimha Rao