NagiosGrapher: HTTP Response Time

by Mike on April 13, 2010

in Nagios

Graph Set Up in NagiosGrapher
The directory /etc/nagios/ngraph.d contains the graph files that will build your graphs.  By default this directory contains subdirectories, templates and in side templates; standard and extra.  The first thing you should do is review the template files in /etc/nagios/ngrpah.d/templates/standard.   This assumes you have  NagiosGrapher already set up.   You will see many of the templates end in .”disabled” which you will then need to move them so they end in a .ncfg”, for example:

mv  check_mysql.ncfg_disabled  check_mysql.ncfg

Check that nagios owns the templates so they can be read.

chown -R nagios:nagios /etc/nagios/ngraph.d

Live Nagios Training is available as well as Linux consulting if you need additional help.

When you start setting up graphs you will find that whether you have used basic checks, nrpe, snmp, ssh or passive checks, you will find some incompatibilities with NagiosGrapher due to the regular expressions (perl) used in the program.  The service_name creates a regular expression which is used by NagiosGrapher to determine the service.    The implications are that you can use a text string to include all of the services that are related.  For example, when you are testing disk usage you could use “disk” in the

The graph output can be modified.  Note the width, height and refresh rates are all able to be changed. The host can be selected and the service is also a selection.

“Average values” is the default but you can select minimum or maximum values as well.

There are 5 graphs available for each service:

These five graphs will allow you to view a history that can provide a useful comparison to current values.

One of the keys to getting these graphs to work correctly is to avoid service_names or graph_values that are the same in two templates.   When you have these similar values you will see that the image for the graph will not appear and in fact, NagiosGrapher will not work correctly.

Graphing HTTP Response Time
There is a template for graphing the response time of your web server.  In order to set this template up you must configure 3 things; set up the host, the service and the template.  The three are related and must be able to communicate based on similar features.  For example, there are two text strings which are important.  The first text string is the check_command string “check_http”.  The second is the service_description “HTTP”.   You can see that these are part of each of the configuration files listed.

define host{
use                     generic-host
host_name               web
alias                   Web Server
check_command           check_http

One thing to note is the service_description which NagiosGrapher will pick up and use.

define service{
use                   generic-service
host_name             web
service_description   HTTP
check_command         check_http

NagiosGrapher Template for check_http

define ngraph{
service_name            HTTP
graph_perf_regex        time=\s*([0-9\.,]+)
graph_value             http_res
graph_units             seconds
graph_legend            HTTP response time
rrd_plottype            AREA
rrd_color               c0c0ff

define ngraph{
service_name            HTTP
type                    CDEF
graph_value             http_cdef
graph_calc              http_res
rrd_plottype            LINE1
rrd_color               000000

# [EOF]

Previous post:

Next post: