0

Hello guys and happy new year!

I am looking for a reliable way to compare WordPress site performance on several different hosters. Key factors for me are:

  • try not to use 3rd party tools or plugins
  • stick to lowest possible level tools (curl, wget etc)
  • test default WordPress installation (no extra plugins or improvements [e.g. CloudFlare])

Bellow is what I plan to check for now.

Measure how long does it take to:

  1. Open default landing page (curl -s http://example.net)
  2. Create and publish a new page (I've written a basic test.php file that basically calls Wordpress functions wp_insert_post and wp_trash_post).

Can you recommend some other approach? Maybe do some stuff using the WordPress API?

Thanks!

Improve this question

2 Answers 2

Reset to default
1

IMHO Your attempt is pointless. On shared hosting you are not going to get consistent results as the results will depend on load generated by other sites on the server, or other sites using the DB server. even worse, the results might change after you have done the testing if new sites were added to the server.

It make more sense to compare true VPS providers, but here you probably do not need anything special for wordpress and can just use whatever are the canonical performance tests for disk access, Apache and mysql.

Improve this answer
3
  • Hi Mark, Case you are describing is true for cheap VPS hosts as well, because a lot of companies oversell VPS resources as well.
    – Radoslav Stefanov
    Commented Jan 8, 2018 at 16:54
  • if you can oversell a VPS, AFAIK it is not a "true" VPS. The only difference that you should have with an empty and full server is network congestion, but internally in the server, there should be no difference at all
    – Mark Kaplun
    Commented Jan 8, 2018 at 18:55
  • .... and if you can oversell, than it is shared hosting with all the problems of trying to get any consistent performance measurements.
    – Mark Kaplun
    Commented Jan 8, 2018 at 18:58
0

Mark Kaplun's comments are true, but you may be looking for less than perfect accuracy. If what you have now is only a guess, or a reliance on marketing materials, same thing, then some testing may provide you with enough information to help you choose between hosts. Remember though that raw performance is only one aspect of why one host might be preferable to another, there's still support and reliability and features to consider, to say nothing of cost.

Anyway, you could set up a simple site with a couple different kinds of pages (one with a post with a few photos, one with a couple database calls, etc) and the use a site like GTMetrix, or Pingdom to measure the performance. You'll want to run the test many times over different times of the day and average them. Make sure that all tests use the same browser and same test server region to keep the tests as similar as you can make them.

This is far from perfect, but much better than nothing.

Improve this answer
1
  • Hi and thanks for the reply! Not sure why some people thing that such test is inaccurate. Measuring performance with simple scripts with API calls and using logs as confirmation is something I have been doing for years now. I like you idea of setting up different pages along with doing the tests during different times of the day, thanks for that!
    – Radoslav Stefanov
    Commented Jan 15, 2018 at 9:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.