Did you know that Apache has a benchmarking tool for testing the HTTP server? It is called ab, and it is pretty great!
As your site grows in popularity, complexity, or size, you will want to test the site to see how it preforms. Having the site crash or lock up during peak time can be devastating for a small blog or e-commerce site. It means lost revenue, and can leave a visitor with a bad impression of your site. This can drive them to generate a bad referral, or worse, go to your competitor.
Users aren't going to care if your site is being bombarded with traffic constantly, they really only care about their own personal experience.
Studies and analytics show that the slower your site is, the impact on your sales or target actions is affected exponentially.
ab tool can be found in most default Apache (httpd 2.2 and 2.4) setups. If you don't have it, you can install the
apache2-utils package and get it from there.
I found a great article on this site that explains a lot of details about using ab, setting up Apache, configuring PHP, and information about the results. You should check it out.
I used the following script, based on that article above, that tests a site in succession and prints the results to a file.
To run the script, download it and unzip it. Then run
chmod +x bench.sh to allow the script to run. Then you can use
./bench and the script will begin.
The article I linked to above mentioned using a delay of 4 minutes between running
Be careful using
ab, as it essentially emulates a DDOS attack, in that it generates as many requests as possible as fast as possible. There is no delay option in ab, so there is no way to emulate something like "10 hits every 10 seconds" or anything like that.
Here are the things you are going to want to pay attention to. These definitions are taken from the ab site.
- Failed requests: The number of requests that were considered a failure. If the number is greater than zero, another line will be printed showing the number of requests that failed due to connecting, reading, incorrect content length, or exceptions.
- Non-2xx responses: The number of responses that were not in the 200 series of response codes. If all responses were 200, this field is not printed.
- Requests per second: This is the number of requests per second. This value is the result of dividing the number of requests by the total time taken.
- Time per request: The average time spent per request. The first value is calculated with the formula (concurrency * timetaken * 1000 / done) while the second value is calculated with the formula (timetaken * 1000 / done).
Failures and errors are generally not good. You should check your logs for anything that happened during your requests.
For Requests Per Second, this tells you how quickly your site was able to process all those requests. Higher is better because it means that your site was able to serve the content without hiccups.
Real World Results
This was helpful after setting up Varnish for this site. I wanted to get an idea of what kind of difference it was making. Well, I believe that I saw about a 500% increase in request per second. This is pretty sweet!