HTTP service uptime nagios plugin

gistfile1.rb

#!/usr/bin/env ruby

# Tweakable knobs

# Which statuses constitute a success?
acceptable_statuses = ["200", "404", "403"]

# How much past history do we consider for each run, in seconds?
history = 60

# Which file are we monitoring?
logfile = './log.txt'

# What are our alert thresholds?
thresholds = {
	:warn 		=> 0.99999,
	:critical => 0.9999
}

begin
	require 'time'
	require 'pp'

	now = Time.new
	log = File.read(logfile).split("\n")
	statuses = {}
	results = {:good => 0, :bad => 0}

	# Used for testing, remove for prodution
	now = Time.parse "2014-05-27 10:00:30 UTC" 

	# Find relevant lines, as defined by being less than one minute old.
	log.reject! do |line|
		# This only works if the date is the first field.
		line_time = Time.parse line

		# We are rejecting lines that fall out of our window
		(now - line_time) > history || line_time > now
	end

	# Parse logs to retreive HTTP status
	# This is made very easy because the status is the last field
	log.each do |line|
		status = line.split(' ')[-1]
		statuses[status] ||= 0
		statuses[status] += 1
	end

	statuses.each do |status, count|
		if acceptable_statuses.include? status
			results[:good] += count
		else
			results[:bad] += count
		end
	end

	result = results[:good] / (results[:good] + results[:bad] + 0.0)
	puts "Service has #{result * 100} % uptime"

	# Generate the exit code.
	exit(0) if result >= thresholds[:warn]
	exit(1) if result >= thresholds[:critical]
	exit(2) if result < thresholds[:critical]

rescue StandardError => e
	puts "Error in script: #{e.message}"
	puts e.backtrace
	exit(3)
end