Dec 26, 2010

Xmas 2010 discounts

For approximately 2 weeks we have 20% discount for all WEBO Site SpeedUp products (licenses). You can get them here:
WEBO Site SpeedUp Premium Edition (save $70).
WEBO Site SpeedUp Lite Edition (save $18).

UPD: action is closed

Dec 8, 2010

v1.3.2 released

It's the last public update for the leading web performance product - WEBO Site SpeedUp - this year. It includes a lot of new features and minor fixes.

  • Improved gzip. There were tested few new gzip-related patterns and WEBO Site SpeedUp support for missed headers, or firewalls, or improper server side configuration, or anything else was improved.
  • Added charset option. Now with charset option (it is available in System Status - Parameters) Extreme caching mode ic completely safe, and no charset will be broken. Charset is detected automatically on the first install, and can be manually changed.
  • Added option to choose between files exclusion or inclusion into CSS Sprites. Very useful option if you have only a few images that must be combined to HTML Sprites, then you can easily add them all instead of adding new and new exclusions.
  • Added option for load time counter (via GA). With this amazing option you can simply count through Google Analytics how many page views you lost with slow website. To perform this you need only to enable WEBO Site SpeedUp with only this counter for a week, and then perform acceleration (with counter enabled). Difference will be clear. All page load time values are being collected into Events (WEBO Site SpeedUp section), and of course you need Google Analytics counter installed on your website.
  • Improved SEO compatibility. Now all combined .php and cached fiels are closed via nofollow and robots.txt from search engines indexation.
  • Improved HTML Sprites.
  • Improved HTML caching.
  • Drupal 7.x added to supported systems.
  • Minor fixes

You can update to the last version from the WEBO Site SpeedUp itself or download it from WEBO Software website.

Special Xmas offer from LoadImpact

Load Impact makes stress test even easier with 50% Xmas discount. You can get it here:

Or just use the following campaign code: XMAS

Dec 2, 2010

Website stress testing

It is important before Xmas season to know more about your website reliability. And of course the exact number of your customers what can you website serve.

Long time there were no tools to perform such an evaluation (except the direct traffic punch) but now you can analyze your website metrics and perform all necessary actions before it go down due to large amount of your clients. One of the best of such tools is Load Impact. Let's look inside this tool and see how simple can be your website bottlenecks analysis.

Configuring a test

Setuping a test. Step 1Of course most of features are available even with free website test. But to get the maximum details and perfect result you should register and use any of paid plans (they start from $9 per day or $39 per month, it's really cheap). So the first thing is to setup the test. You can use the only website home page URL to get almost all data about its reliability. But can get a bit more.

With Proxy recorder (under Advanced mode -> Load script generation -> Record session) you can setup a set of requests for your website (with various delays); this will completely emulate your visitors' behavior and can tell you how many simultaneous visits your website can serve.

Setuping a test. Step 2: configuring proxyProxy recorder can be configured in a few steps (you just need to enable a proxy on a given port inside your browser, Firefox in this case), then start recording and go through the target website (to collect all requests, even dynamic and AJAX ones). Then just stop recording and get all these requests in textarea.

Setuping a test. Step 2: verifying proxyPlease be sure that proxy is configured correctly and test proxy settings before starting the test.

Setuping a test. Step 3: choosing stepsBefore you finish with test setup you should choose amount of visitors to emulate. For a small website 50 visitors will be quite enough.

If you just want to test website reliability for 500-1000 visitors, you can setup step as 100 (not to overload the system and overhead traffic limit).

In the case if you need exact picture of server down due to high load you need to choose 10-20 visitors in step and get complete picture. But this can take several hours to be completed.

Before you can save the test you need to verify it. Usually it is just a formal procedure to be sure that test is correct.

Running a test

After you have created the proper configuration you can run it. Please be aware of long duration of test performance - it can take up to several hours. After test is completed you will get a lot of different charts about pages and resources performance. Let's consider them more.

Test results. Acc. user load timeThe main graphics are user load time and accumulated user time. The first will tell you the delay for users before they get HTML page under current amount of simultaneous users online. The second will tell you how website is working under load, how long its pages are being loaded (totally, server and client side), and where is the limit of its reliability (where time becomes more than 10-15 seconds).

Also you can simply compare how different type of resources are being loaded - to find bottlenecks and to fix them properly. For example you can have relatively fast HTML document, but a lot of expenses on dynamic images generation. Or main expenses will be concentrated in HTML latency only. This will appear only after the test is completed.


Load Impact is a very easy tool to perform such tests, and simple load tests are completely free. So you can quickly check your website reliability and be sure that you are ready for Xmas shopping season :)