Their sub-title for this session is “It’s not exactly socket science.” Love it!!!
Today’s session is about automated web performance testing before 5pm. It’s the last session of the day and I’m hoping to get out of here before 5:30pm. These guys are from Neustar. The examples can be found for this workshop here.
Looking for http issues, but also looking for responsiveness. Why automate?
Going to be using a primary set of tools (Selenium, Watir, QTP and Silk Performer). Selenium is the focus. Plan on using Browser Mob Proxy and a HAR viewer.
Note to self: I think it’s important for everyone on the team to learn Selenium programming. There’s really no reason that each person on the team shouldn’t be sufficient at writing their own Selenium code. This is something holding back the SPEs in my opinion. They are dependent on their SPE pairs to write their Selenium scripts on their behalf. I also think it’s important for the SPTs to learn Selenium as well so that they can take advantage of browser performance testing on their own as well.
HTTP archive with simple format (JSON/JSONP), tons of data, easily extensible and now becoming the standard.
Tools for storing HAR format in demo is called Harpoon.
- Guice (Google)
Can access at http://labs.webmetrics.com:8080
Note to Self: This is actually a pretty cool idea for us to consider integrating with Galileo. The standard is the HAR format, which we could easily persist and prepare views within Galileo for visualization purposes. Also note that dynaTrace offers a HAR viewer.
Need to do some research on Web Timings specification. This happened in the last year. Surprised I didn’t see this. Here’s an interesting blog about it from Microsoft’s perspective. There is apparently a Web Performance Workshop Group which kind of surprised me. There are apparently 2 other specs to look at” Resource Timing and User Timing.
It’s an open source tool and saves HTTP archive. Also has a native Java API. Can use REST API for other languages. I’m curious if this would solve our grid problem in the lab. Could we set up a browser mob proxy farm in our lab?
Hmmm…it looks like all it’s really doing is capturing the HAR.
Note: Selenium saves screenshots. Totally didn’t know that. Great for debugging purposes.
- http://www.slideshare.net/watsonmw/performance-monitoring-in-a-day: This is the same presentation, but in greater detail
- http://eurosys2011.cs.uni-salzburg.at/pdf/postersubmission/eurosys11-posteranddemo-cecchet.pdf: Worth reading!!!