check_nginx_status.pl

Project Notes

CHECK_NGINX_STATUS
==================

Nagios check checking the nginx_status page report from nginx. Tracking Active connections processes, request per second,
connections per seconds, Connections status.

Nginx Monitor for Nagios version 0.10
GPL licence, (c)2012 Leroy Regis

Known Bugs:
===========

You may need to adapt the nagios libraries path on the script (currently /usr/local/nagios/libexec), line 18 of the script. It as been reported on different places such as /usr/local/libexec/nagios or /usr/share/nagios/libexec.

Graphing Results
================

This check is compatible with performance data analysis for passive monitoring (RRD based graphs). So the script output is available for graph tools.
@maethor made a PNP4Nagios template available in this Gist:
* https://gist.github.com/maethor/8714514

Script Documentation:
=====================

Usage: ./check_nginx_status.pl -H <host ip> [-p <port>] [-s servername] [-t <timeout>] [-w <WARN_THRESOLD> -c <CRIT_THRESOLD>] [-V] [-d] [-u <url>] [-U user -P pass -r realm]
-h, --help
print this help message
-H, --hostname=HOST
name or IP address of host to check
-p, --port=PORT
Http port
-u, --url=URL
Specific URL to use, instead of the default "http://<hostname or IP>/nginx_status"
-s, --servername=SERVERNAME
ServerName, (host header of HTTP request) use it if you specified an IP in -H to match the good Virtualhost in your target
-S, --ssl
Wether we should use HTTPS instead of HTTP
-U, --user=user
Username for basic auth
-P, --pass=PASS
Password for basic auth
-r, --realm=REALM
Realm for basic auth
-d, --debug
Debug mode (show http request response)
-m, --maxreach=MAX
Number of max processes reached (since last check) that should trigger an alert
-t, --timeout=INTEGER
timeout in seconds (Default: 15)
-w, --warn=ACTIVE_CONN,REQ_PER_SEC,CONN_PER_SEC
number of active connections, ReqPerSec or ConnPerSec that will cause a WARNING
-1 for no warning
-c, --critical=ACTIVE_CONN,REQ_PER_SEC,CONN_PER_SEC
number of active connections, ReqPerSec or ConnPerSec that will cause a CRITICAL
-1 for no CRITICAL
-V, --version
prints version number

Note :
3 items can be managed on this check, this is why -w and -c parameters are using 3 values thresolds
- ACTIVE_CONN: Number of all opened connections, including connections to backends
- REQ_PER_SEC: Average number of request per second between this check and the previous one
- CONN_PER_SEC: Average number of connections per second between this check and the previous one

Examples:

This one will generate WARNING and CRITICIAL alerts if you reach 10 000 or 20 000 active connection; or
100 or 200 request per second; or 200 or 300 connections per second
check_nginx_status.pl -H 10.0.0.10 -u /foo/nginx_status -s mydomain.example.com -t 8 -w 10000,100,200 -c 20000,200,300

this will generate WARNING and CRITICAL alerts only on the number of active connections (with low numbers for nginx)
check_nginx_status.pl -H 10.0.0.10 -s mydomain.example.com -t 8 -w 10,-1,-1 -c 20,-1,-1

theses two equivalents will not generate any alert (if the nginx_status page is reachable) but could be used for graphics
check_nginx_status.pl -H 10.0.0.10 -s mydomain.example.com -w -1,-1,-1 -c -1,-1,-1
check_nginx_status.pl -H 10.0.0.10 -s mydomain.example.com

Licence: GNU GPL v3