What are digital ocean stress tests for a wordpress website?

A

Artur Bekerov2014-02-02 10:25:23

Solid State Drives

Artur Bekerov, 2014-02-02 10:25:23

Hello.
I found several services for stress testing a website on WordPress.
Standard installation and several pages.
Default configured lamp
Surprised by the results.
A 512 MB virtual machine holds only 50 simultaneous visitors and 1-2 requests per second on the loadimpact.com service . I don't understand if it's too much or too little?
after that, the muscle falls and that's it.
What options can be used to make everything work fine. I want to get away from shared hosting (unprofitable already).
Is it correct to auto-restart the muscle?
Fine-tuning for Apache for a 512 MB virtual machine on an SSD?
What services of an adequate assessment can you know?
Still looked before the fall.
SBU Total number of simulated clients (VU or SBU) active. 41
Connections active 82 Number of open TCP connections to target system.
Requests 1045 (7 req/s) Current number of requests per second.

Reply

Answer the question

In order to leave comments, you need to log in

6 answer(s)

V

Vlad Zhivotnev, 2014-02-02
@inkvizitor68sl

Hire an admin)
And according to the subject - if WordPress is not clogged with unsuccessful plugins (well, its database is less than 250 mb), then on this virtual machine you can keep 5-7 requests per second. If you work hard - then 10.
"The number of visitors to the site" is not an indicator of anything at all.

P

Puma Thailand, 2014-02-02
@opium

How often do you have 50 visitors on your site at one time?
Any shared has clearly more than 512MB of RAM.

S

Sergey, 2014-02-02
@bondbig

The muscle falls, most likely from a lack of RAM. Need to tune in. For a start, you can take this as a sample (such a config on my experimental droplet in 512Mb):

[client]
port		= 3306
socket		= /var/run/mysqld/mysqld.sock
[mysqld_safe]
socket		= /var/run/mysqld/mysqld.sock
nice		= 0
[mysqld]
user		= mysql
pid-file	= /var/run/mysqld/mysqld.pid
socket		= /var/run/mysqld/mysqld.sock
port		= 3306
basedir		= /usr
datadir		= /var/lib/mysql
tmpdir		= /tmp
lc-messages-dir	= /usr/share/mysql
skip-external-locking
local-infile=0
bind-address		= 127.0.0.1
key_buffer		= 16M
max_allowed_packet	= 16M
thread_stack		= 192K
thread_cache_size       = 8
myisam-recover         = BACKUP
query_cache_limit	= 1M
query_cache_size        = 16M
log_error = /var/log/mysql/error.log
expire_logs_days	= 10
max_binlog_size         = 100M
[mysqldump]
quick
quote-names
max_allowed_packet	= 16M
[mysql]
[isamchk]
key_buffer		= 16M
!includedir /etc/mysql/conf.d/

+ for stability, it's better to create a swapfile
Well, then - standard procedures:
1) put the server on monitoring (if it's simpler, then munin / monitorix)
2) load
3) watch logs and monitor
4) tune service and CMS configs
5) GOTO 2 up to until an acceptable result is achieved.
And from online services, it is better to load using blitz.io (if we are talking about free versions), in the "free" version it generates a more serious load than loadimpact.

D

Dan Ivanov, 2014-02-02
@ptchol

Buy in the same DO for each component of the system 1 instance.
1 mysql, 1 php-fpm, 1 nginx. drive it into the same newrelic at a free rate.
15 bucks and you'll be happy.
In the future, it will be easier to scale horizontally first.

D

DuD, 2014-07-23
@DuD

I'm willing to bet that putting nginx in front of apache will get at least +10-20 online.