📜 ⬆️ ⬇️

How much does habra effect in Jelastic

Hello!
Last year, when Jelastic started to support PHP in Rusonyx, I first installed CMS 1C-Bitrix on my test environment.
And the idea to write a post about it for a long time spun in my head. But to write about installing CMS, even for such an interesting and new platform as Jelastic, does not at all pull to the post. Therefore, an idea emerged - to check the work of the online store based on Bitrix at peak loads in combat mode, and also to find out one of the most burning questions - how much will it cost?



to take part in habraeffekt: bitrix.j.rsnx.ru
')


Above the cut is a link to our experienced online furniture store, the cost of which we find out the cost. Be kind - click on the links and sections on this site.

Preliminary preparations


For testing, we created the Jelastic environment with the following components:
- Apache Web server:
min: 1 cloudlet (128MB, 200MHz)
max: 32 cloudlet (4GB, 6400MHz)

- MySQL server
min: 1 cloudlet (128MB, 200MHz)
max: 32 cloudlet (4GB, 6400MHz)



First of all, you need to install Bitrix in our environment, this is done through a universal installer script, which can be downloaded from the official site , then run it at yousite.com/bitrixsetup.php and follow the instructions of the installer:


After that, in accordance with the requirements of the CMS, you need to configure php.ini (everything is done simply and naturally):


These are the values ​​that were set when the site was empty (Mysql ate memory for 4 cloudsets, Apache takes a minimum):


We start the first testing


Testing with Apache Benchmark is already beaten and uninteresting, because we have taken the loadimpact.com paid service, which can load the Web site dynamically, from different points of the world and emulates the work of real browsers.
The loading scenario was as follows: the number of virtual visitors will grow from 1 to 600 within 20 minutes.


Pre-test measurement of money spent over the past day (17 rubles spent):


Three ... Two ... One ... Run!
During the test, the number of klaudletov increased almost to a peak (27 of 32):


As soon as the load subsided, the resource consumption dropped to the initial level - the scaling worked!

The load charts were as follows:


At the end of the test, you can measure the money spent:


Total: 46.65 rubles for a 20 minute load of 600 pseudo-users.
Not so much, right?

A spoon of tar


All this is of course good, but I do not have confidence in such a test. Too many unknowns - how it loads the site, what it does at this moment, what its actions are with JS, etc. etc. Also a solid problem is the fact that all participating instances are located anywhere, but not in Russia.
Because these figures in fact did not say anything to me - it scales - yes, money is being spent - yes, so what?
After all, each user of cloud platforms is interested in - what wave of visitors to his project will fly into?
Because I have a suggestion!

And let's arrange a website bitrix.j.rsnx.ru habraeffekt!


I will be very grateful to everyone who visits our furniture test store bitrix.j.rsnx.ru and walk a little bit through its pages.
And tomorrow at 13-00 we will stop testing, after which we will post an update and tell you how much our habraeffection cost.

Let's try to guess how much money will eat Jelastic per day? We'll give the most accurate Jelastic currency to the most accurate one, how much habraeffect will be spent per day!

RESULTS:
Alas, there was no habraeffect: (in two days only 2115 unique IPs came to our online store.
Yesterday and today Jelastic ate 117r

Charts from the moment of publication and the beginning of the experiment:


Although the experiment failed, Agel_Nash won and gets a free test :) Write in a personal!

Thanks to all! See you later!

Source: https://habr.com/ru/post/175639/


All Articles