Nginx statistics in cacti

Based on the previous article , only this time we collect statistics for the nginx web server .


First you need to install nginx with the module http_stub_status_module . And turn it on, for this we add the lines in /usr/local/etc/nginx/nginx.conf (for freebsd):

location / nginx_status {
   stub_status on;
   # disable access_log if requared
   access_log off;
   #allow XX.YY.AA.ZZ; better resolved only for 127.0.0.1
   #deny all;
}

Now, when requesting http: // localhost / nginx_status, the server will output something like this:

Active connections: 1
server accepts handled requests
2 2 4
Reading: 0 Writing: 1 Waiting: 0

Now download the script for Cacti from here: forums.cacti.net/download/file.php?id=12676 "href =" forums.cacti.net/download/file.php?id=12676 Unzip it

and put it in / usr / local / share / cacti / scripts / files get_nginx_clients_status.pl and get_nginx_socket_status.pl.

Now open cacti, go to Import Templates and import cacti_graph_template_nginx_clients_stat.xml and cacti_graph_template_nginx_sockets_stat.xml

There should be 2 new templates for graphs:

Nginx_clients_stat
Nginx_sockets_stat

We create graphs, prescribe the URL from which statistics will be taken and observe them =)



(the graphs are not mine)


PS the PERL LWP :: UserAgent module is required, it is put simply:

perl -MCPAN -e 'install "LWP :: UserAgent"'

Based on: http://forums.cacti.net/about26458.html
A similar article for lighttpd: habrahabr.ru/blogs/sysadm/66602

UPD :
Legend to the first schedule :
Active connections - how many clients are served.
Reading - how many connections are in read state.
Writing - how many connections are in the recording state.
Waiting - keep-alive connections or in a state of processing a request.

Legend to the second schedule :
server accepts - how many connections were accepted;
handled- how many of them were processed, but not closed immediately, the connection immediately closes if the connection table is full.
requests - how many requests are served. With keep-alive, there can be multiple requests on the same connection