Writing a Nagios Plugin in Python

Our Solarwinds Network Performance Monitor has a problem rendering custom reports on occasion. For something like that, there isn't an existing plugin for Nagios. Writing these plugins is easy. All there is to it is exit statuses. After reading this, you should have an idea of how to write a Nagios plugin for a variety of web applications.

#!/usr/bin/env python

from mechanize import Browser
from optparse import OptionParser

# Exit statuses recognized by Nagios
UNKNOWN = -1
OK = 0
WARNING = 1
CRITICAL = 2

def open_url(br, url):
    """Use a given mechanize.Browser to open url.

    If an exception is raised, then exit with CRITICAL status for Nagios.
    """
    try:
        response = br.open(url)
    except Exception, e:
        # Catching all exceptions is usually a bad idea.  We want to catch
        # them all to report to Nagios here.
        print 'CRITICAL - Could not reach page at %s: %s' % (url, e)
        raise SystemExit, CRITICAL
    return response

# I'm going to be using optparse.OptionParser from now on.  It makes
# command-line args a breeze.
parser = OptionParser()
parser.add_option('-H', '--hostname', dest='hostname')
parser.add_option('-u', '--username', dest='username')
parser.add_option('-p', '--password', dest='password')
parser.add_option('-r', '--report_url', dest='url',
    help="""Path to report relative to root, like
    /NetPerfMon/Report.asp?Report=Hostname+__+IPs""")
parser.add_option('-v', '--verbose', dest='verbose', action='store_true',
    default=False)
parser.add_option('-q', '--quiet', dest='verbose', action='store_false')

options, args = parser.parse_args()

# Check for required options
for option in ('hostname', 'username', 'password', 'url'):
    if not getattr(options, option):
        print 'CRITICAL - %s not specified' % option.capitalize()
        raise SystemExit, CRITICAL

# Go to the report and get a login page
br = Browser()
report_url = 'https://%s%s' % (options.hostname, options.url)
open_url(br, report_url)
br.select_form('aspnetForm')

# Solarwinds has interesting field names
# Maybe something with asp.net
br['ctl00$ContentPlaceHolder1$Username'] = options.username
br['ctl00$ContentPlaceHolder1$Password'] = options.password

# Attempt to login.  If we can't, tell Nagios.
try:
    report = br.submit()
except Exception, e:
    print 'CRITICAL - Error logging in: e' % e
    raise SystemExit, CRITICAL

report_html = report.read()
# class=Property occurs in every cell in a Solarwinds report.  If it's not
# there, something is wrong.
if 'class=Property' not in report_html:
    print 'CRITICAL - Report at %s is down' % report_url
    raise SystemExit, CRITICAL

# If we got this far, let's tell Nagios the report is okay.
print 'OK - Report at %s is up' % report_url
raise SystemExit, OK

To use our plugin, we need to do a bit of Nagios configuration. First, we need to define a command.

define command{
    command_name    check_npm_reports
    command_line    /usr/local/bin/reportmonitor.py -H $HOSTADDRESS$ $ARG1$
}

After that, we define a service.

define service{
    use         generic-service
    host_name           solarwinds-server
    service_description Solarwinds reports
    check_command       check_npm_reports!-u nagios -p some_password -r '/NetPerfMon/Report.asp?Report=Hostname+__+IPs'
}

Posted by Tyler Lesmann on September 3, 2009 at 13:37

Tagged as: mechanize nagios optparse python screen_scraping