Top Web Performance Testing Articles


Apica´s Web Performance Blog is your source for tips, trends and tools for optimizing your web performance. Track current web performance metrics and benchmarks and new technology in web performance testing.

[New Feature Release] Apica Load Test Portal Supports ZebraTester Version 5.5A

Apica Load Test Portal Supports ZebraTester Version 5.5A


ZebraTester version 5.5A offers an improved development of performance test scripts and will be released shortly. Below, we’ve got some examples of how this feature will help you.

3 new Types of Variable Extractors
The following 3 new types of Variable Extractors have been added and will support random extraction.

  • Left and Right Boundary based Variable Extractors
  • RegEx based Variable Extractors
  • XPath based Variable Extractors

 

Extract variable by left and right boundaries

Extract variables by a Left and Right Boundaries

Extract variables by using Regular Expressions

Extract variables by using Regular Expressions

Extract variables by using XPath

Extract variables by using XPath

Example

Example

An extract of HTML response content using Left and Right Boundaries
In this example, a date parameter was extracted.

For more information on ZebraTester 5.5A please see here.

 

Request Download Time Breakdown

 

When analyzing performance test results it can be necessary to go down to the details.
We have added more details for the page download time and can now present the following seven metrics for each request.

DNS Resolve Time – The average time it takes to resolve the domain/subdomain to an IP address

TCP Socket Connect Time
– The average time it takes from the client to make a network connection to the target server.

SSL Handshake Time – The average time it takes to setup encryption (3 TCP packets).

Request Transit Time – The average time it takes for all bytes to be sent to the target server from when the network connection is established.

Response Header Wait Time – The average time it takes for the target server to send the first byte from when the request was sent.

Response Header Receive Time –The average time it takes for the server to send the response header from when the first response byte was received.

Response Content Receive Time – The average time it takes to receive the response content data from when the header was received.

page-view

The information can be found in the Live View Page while running the test.

 

The metrics are also available on the Test Result Page under Page & URLs

The metrics are also available on the Test Result Page under Page & URLs

 

Note that depending on the type of request and the test scenario, not all metrics has values for every request.
DNS lookup, for example, is usually done in the first request of the test.

 

Saving Non-Functional Requirements

 

Non-functional requirements are an important part of performance tests.
With this new feature, users can add response time requirements for all Pages and Transaction in a test. Page and Transaction names come from the ZebraTester script used in the test.
The requirements can then be compared to latest and previous test results on the Continuous Integration page.

To enter Non-functional requirements to a test start by clicking on the new green icon on the Scenario Manager page and the uploaded ZebraTester script related to the test.

Scenario-manager

Edit-non-functional-requirements
This opens a new dialogue where the user can enter a Non-functional Requirement for each page in the test. Page names are extracted from the ZebraTester script. Values are entered in milliseconds for each page.

 

Relating ZebraTester Script Uploads with Apica Load Test Projects


ZebraTester scripts are now related to Projects when uploaded to the Portal. This is the first step in making Project a placeholder for everything relatable to a test. ZebraTester scripts are uploaded from the Scenario Manager page as before.
When selecting ZebraTester file(s) to upload, the user can now relate ZebraTester scripts to existing projects or create a new project for the uploaded script(s).

Relating-ZT-scripts

 

Performance Monitoring of Servers Without Showing the Hostname 

 

Sometimes server host names are considered sensitive information.
This feature improvement enables the user to hide the host name from the PMA agent in the Apica Load Test Portal.

perf-monitoring

To hide the hostname add the following configuration in the PMA agent configuration file.

[client.conf]
hideLocalNetworkInfo=”true”

 

Page and Transaction Response Times in Live View

 

When running performance tests on a new or altered system, it’s extremely helpful to see how the new release performs in real-time to detect real-time changes. Detecting a change early in a test will save a lot of time. The whole test does not have to be completed.

We have added 2 new Graphs in the Live View for this purpose.
The new graphs show page and transaction response time in relation to number of concurrent users.

live-view

Ability to Compare CI Test Results to Non-Functional Requirements


We have extended the Continuous Integration result page with non-functional requirements.
This feature allows the user to compare test results both to previous tests and to the Non-functional requirements.

When the user opens the Continuous Integration result page and the test selected contains Non-functional requirements the graphs will have the Non-functional requirements included.

non-fucntional-metrics

The Non-functional metrics is showed as a red trend in the top graph and as a red bar for each page. The user can click on each page to change the top graph view to a specific page and compare how the application performs in relation to the Non-functional requirements.

 

Real-Time Settings with Delay Per User

 

You may want to decrease the load and slowly ramp it up again.
Therefore we have now added a new delay parameter in the Real Time Settings.
non-fucntional-metrics

In the Real Time dialog, a user can now change the number of users, ramp-up delay, and duration.

Here is an example of a test where the load was first ramped-up to 30 then ramped-down to 1 user and finally slowly ramped-up to 30 again.

Here is an example of a test where the load was first ramped-up to 30 then ramped down to 1 user and finally slowly ramped-up to 30 again.

 

Print and Export Graphs to File


A print and export function is added to the page and transaction response time graphs on the Live View page, and on the Result Page.

 

New Call-to-action