QABot

[qabot-blog]

Traditional website monitoring is becoming irrelevant

"Up and Running" - what does it really mean for your website? How do you know that your web app is really up? Is ping enough? Does "HTTP 200 OK" mean it's valid? Only you can answer these questions, it's up to you to decide what's best for your app. But I do believe that a better way to monitor your website is to make sure that users see what you want them to:

UI Components are displayed correctly (meta tags, buttons, any html tag or text on page)

Ajax and Javascript content is loaded properly

Web resources can be opened

3rd-party components are working as expected (Disqus, Facebook, Twitter, etc...)

All these cannot be monitored using traditional approach, the only way you can be sure that Facebook dialog is still opening is by actually load your page in your web browser and see for yourself. But we cannot do this everyday right? In a same way you want to make sure that your product list is loading fine and not timing out, even when you've got 100k of items in stock, although initially had only 1k and it was fast.

We know web browsers are being used in manual and automated testing, but why can't we use it for monitoring production version? If we're getting value out of regression testing, why can't we get value out of running the same tests, or most of them, against production environment? This is the real "Up and Running" in my mind.

Thank you.

[qabot-blog]

Use Selenium to monitor web applications

The are no doubts about benefits of regression and automated testing. Lately browser automated tests are becoming even more popular and one of the leaders in this area is Selenium (seleniumhq.org), an open-source testing tool that's using real web browsers to perform your tests.

Here is a simple JUnit test that is checking home page's title.

public class Example1Test { private WebDriver wd; @Before public void before() { wd = new FirefoxDriver(); wd.manage().timeouts().implicitlyWait(30, TimeUnit.SECONDS); } @Test public void test_title() { wd.get("http://example.com"); assertEquals("Welcome to exemple.com!", wd.getTitle()); } @After public void after() { wd.close(); wd.quit(); } }

This is a basic Selenium test, but does quite a few things. It will open Firefox browser, navigate to example.com, assert the title and if it doesn't match junit will fail and let you know where. Usually you'd run tests like this in your continuous integration setup, after each build and deployment. This is a good practice, but I think we can go even further and run these tests continuously, as part of continous regression testing setup, against production web apps. How often it's up to you, but the more tests you have the more chances to prevent some issues that occur only in production. One thing to mention is that if you're using Google Analytics, you have to make sure that Selenium tests don't mess up your stats, you can though avoid that*

At QABot, we're running our smoke tests every 15 minutes, which include checking if ToS, Privacy and Contacts pages are available, but we're not stopping here, we'll be adding more tests once we have more functionality and more content.

* - http://support.google.com/analytics/bin/answer.py?hl=en&answer=1034840

[qabot-blog]

A better way to monitor web apps

We believe in Continuous Regression Testing. We are sure that continuous app validation reduces downtime and helps avoid disasters. That's why we created QABot, an automated web app monitoring service to help you create more stable web apps. QABot uses real browsers and gets the monitoring as close as possible to real usage. At it's core, QABot is based on Selenium 2, a popular open source web browser automation tool. 

You define the rules, we check them for you!

How does it work?

The basic idea is that you get started with your Selenium IDE test cases, upload them to QABot, we run them for you and let you know if anything goes wrong!

Please note, even though we covered a vast number of Selenium commands, there may be cases where you'll have to adjust your test case in order to be understood by QABot. Our goal though, is gradually reduce the need of doing that, we're constantly looking into increasing coverage of supported Selenium API.