[00:03] Benchmark your WordPress is a good way to estimate how far it can serve, avoid interruptions to your users in the future.
[00:24] Basically, you use benchmark tools to force your server (WordPress/Apache actually) work at peak (~100%) to see its capacity.
[00:34] "Siege can be defined as an HTTP load testing and benchmarking utility. According to the official definition, it was designed specifically to let web developers measure their code under duress, to see how it will stand up to load on the internet" -- dotlayer.com
[00:53] Fortunately, there are so many free tools to help you done this job, which was introduced on this page.
[1:05] I will use "siege" the simple but powerful command-line tool of Linux.
I'm migrating my WordPress blog from Amazon AWS to Cloudrino (Apache PHP 5.6 to LiteSpeed PHP7, actually);
during the phase of waiting for DNS propagation, I do a little benchmark to see which one is better.
[1:55] Take a look at the SIEGE tool.
[2:03] "The t2.nano is a very useful instance capable of running workloads such as a low-traffic WordPress blog or a JMeter load generator. They are great instances for doing training and some development and testing (depending on the applications you are developing or testing)" -- concurrencylabs.com
[2:14] This simple tool can give you a detail about your web server quickly, and the preparation is quite easy.
Furthermore, it simulates user traffics, so we will have an objective result.
[2:30] Siege is an open source regression test and benchmark utility
[3:11] "When choosing your web server, you have several choices; Apache, Nginx, IIS, Caddy, and Lig
httpd are all popular projects. We will be covering Apache and Nginx in this guide" -- code.tutsplus.com
[3:26] My old back-end lays on Amazon AWS, optimized by Bitnami.
[3:42] This is my new site, which comes from Cloudrino, run on LiteSpeed web server; I built it by myself.
[3:59] "When a web server process, or an interpreter like PHP, don’t have enough RAM, they start swapping, and swapping effectively means using the hard disk to supplement RAM memory. The effect of this is increased latency every time this memory is accessed" -- sitepoint.com
[4:09] "The type of EC2 instance will depend upon the number of instances that you use for apache , (as per my recomendation m3.2xlarge will be a good option ). You can also use Cloud front service that will improve your sites latency" -- quora.com
[4:18] The benchmark procedure is quite simple.
We will install this tool then do a cross-test via SSH connections, from Amazon AWS to Cloudrino and so on.
[5:52] "It’s obvious that separating web and database server into two Amazon instances gives you quite a boost. And it’s obvious that I don’t know much about performance measuring" --http:/mamchenkov.net” target=”_new”>>http://mamchenkov.net_
[6:07] Cloudrino to Amazon AWS.
Currently, the domain minhthang.name.vn is pointing to AWS.
[6:19] "If our search statistics are any indication, the question on everyone’s mind is “How does OpenLiteSpeed compare to NGiNX and Apache?” We think that’s a question that deserves an answer, so we set up a test environment and got to work" --http://openlitespeed.org” target=”_new”>>http://openlitespeed.org
[6:28] Amazon AWS to Cloudrino.
The new/temporary site is running via an IP address 103.250.xxx.xxx
[6:46] Siege will work fine as default.
However, you can put more reality into the tests by tweaking parameters via switches, like the number of concurrent users in this case (-c).
Leave them run for a while, about 20 minutes.
Stop the tests by press Ctrl+C on terminals.
It is recommended that two tests share the same duration.
However, don't worry, the results will be broken down into the rate per second unit.
[7:00] "Amazon Aurora is a drop-in MySQL replacement built and managed by AWS that comes with unparalleled speed, scalability, and durability" -- getshifter.io
[7:07] To get more precise of the benchmark, follow this guideline.
[7:28] Part 2 will help you: solve errors of the benchmark, tweak parameters, understand result's numbers, create config files, etc.
So, be patient!