| || |
HAR (HTTP Archive Format) analysis and tools
is anyone doing anything interesting with HAR files for web pageload performance?
HAR (HTTP Archive), is a file format that can be used by HTTP monitoring tools to export collected data.
The HAR format is based on JSON, and is used by tools that consume/produce data collected by monitoring HTTP communication. These files contain a log of HTTP client/server conversation and can be used for additional analysis of page load performance.
HAR files can be viewed/analyzed with various tools:
I have built HAR capture into my selenium-based test framework, for automated collection.
If anyone is interested, here is some info including a screencast: http://coreygoldberg.blogspot.com/20...ng-python.html
... just wondering what other people are doing around performance analysis using HAR?
Re: HAR (HTTP Archive Format) analysis and tools
Nice. This might be something we could use.
We have a framework for running automated JMeter tests which gives us a jtl file containing data for response times under load that we then store in mysql. This is open source.
We also have a presentation layer for these results (which is currently private but could be released under an open source license soon) and it is here that I am most interested in your project. I'm very keen to have a setup where I can trigger page load tests from the command line and then have the results stored in a location where I can visualise results over time.
If we could extend our visualisation to include data from HAR files, or even start a new project, where page load performance was trended from release to release, it would really add benefit.
Do you have any thoughts on:
1. Extending your script to store data in a DB?
2. Trending HAR data? (There's a lot of cool stuff on presenting single files but not much on comparing multiple files.)